o 
 
 3 , 
 
 The Census Bureau 
 
 A NUMERATOR AND DENOMINATOR 
 FOR MEASURING CHANGE 
 
 Technical Paper 37 
 
 U.S. DEPARTMENT OF COMMERCE 
 
 BUREAU OF THE CENSUS 
 
u 
 
 o 
 
 
 '^6-l9l fe 
 
 Technical Paper 37 
 
 The Census Bureau 
 
 A NUMERATOR AND 
 DENOMINATOR FOR 
 MEASURING CHANGE 
 
 Issued June 1975 
 
 /f°\ 
 
 *''iriso» f ' 
 
 U.S. DEPARTMENT OF COMMERCE 
 
 Rogers C. B. Morton, Secretary 
 
 James L. Pate, Assistant Secretary for Economic Affairs 
 
 o 
 
 S> BUREAU OF THE CENSUS Vincent P. Barabba, Director 
 
BUREAU OF THE CENSUS 
 
 Vincent P. Barabba, Director 
 Robert L. Hagan, Deputy Director 
 
 ACKNOWLEDGMENTS 
 
 These seminar transcripts were edited and prepared by the 
 Director's Office in cooperation with the Publications Services 
 Division of the Social and Economic Statistics Administration. 
 Within the Publications Services Division many individuals made 
 significant contributions in the areas of publications planning and 
 design, editorial review, composition and printing procurement. 
 Jack L. Osborn, Administrative Assistant to the Director was 
 responsible for planning and review of the material. 
 
 The cooperation of the University of Southern California School 
 of Public Administration and Charles Grace, Director of its 
 Program Development, as well as the government and private 
 industry participants, is gratefully acknowledged. 
 
 Library of Congress Card No 75-619192 
 
 SUGGESTED CITATION 
 
 U.S. Bureau of the Census, The Census Bureau, A 
 Numerator and Denominator for Measuring 
 Change, Technical Paper 37, U.S. Government 
 Printing Office, Washington, D.C. 1975 
 
 For sale by Superintendent of Documents, U.S. Government 
 Printing Office, Washington, D.C. 20402. Price $4.25. 
 Stock Number 003-024-010810 
 
CONTENTS 
 
 Page 
 
 Aspects of the Census Bureau — Vincent P. Barabba 1 
 
 Social Observatories. . Through a Glass Less Darkly — Vincent P. Barabba 11 
 
 A Discussion of Privacy - Vincent P. Barabba 17 
 
 The Right of Privacy and the Need to Know — Vincent P. Barabba 23 
 
 Monitoring the Nation's Health — Edward B. Perrin and Doug/as Williams 30 
 
 Measuring National Wealth and Well-Being — Sidney L. Jones 41 
 
 Profiles of Small Urban Area Change - Richard Hanel 53 
 
 The Role of Private Enterprise in Developing Indicators — Richard Wirthlin 77 
 
 The Graphic Presentation of Statistics - Vincent P. Barabba 89 
 
 The Census Function — Daniel B. Levine 104 
 
 Planning the Federal Statistical System— A Historical Perspective — Joseph Duncan . . 138 
 
 Commercial Use of Census Data — Max Eveleth, Jr 151 
 
 The Evolution of Social Indicators — Lee Preston 159 
 
 Social Indicators for Corporate Management — Lee Preston 170 
 
 The Choice of Methodology in Social Problem Solving — Harold Nisselson 177 
 
 GBF/DIME System— The Diffusion of an Innovation — Vincent P. Barabba 189 
 
Digitized by the Internet Archive 
 
 in 2012 with funding from 
 
 LYRASIS Members and Sloan Foundation 
 
 http://archive.org/details/censusbureaunumeOOunit 
 
INTRODUCTION 
 
 The U.S. Census Bureau has been described in a variety of 
 manners. It has been called "a data factory," "the largest statis- 
 tical agency in the western world," "a Byzantine bureaucracy," 
 "the greatest threat to individual liberties," and "one of the best 
 brand names in America." 
 
 All of these phrases indicate that much like the fable of the 
 blind man and the elephant one can discover a variety of 
 "natures" in the activities and products of the Census Bureau. 
 
 From time to time individuals have attempted to describe 
 aspects of the Census Bureau, and some have been more success- 
 ful than others. This publication does not attempt to tell the 
 Census Bureau's story, rather it is designed to discuss some of the 
 roles that accurate and timely information can and should play in 
 the making of meaningful decisions. 
 
 The format of this publication is quite simple. It is the 
 result of 11 experts in the information and data processing fields 
 talking about the uses and potential uses of information to 20 
 Masters and Ph.D. candidates at the University of Southern Cali- 
 fornia's School of Public Administration. 
 
 These participants were a unique group of students, in that 
 the vast majority were full-time employees of State and local 
 government agencies or private industry. Most, if not all of the 
 students were crucially involved with the day-to-day evaluation or 
 dissemination of accurate and timely data in fields as varied as 
 health care, law enforcement, and engineering. It was for this 
 reason that the Bureau thought such a classroom format could 
 produce the type of meaningful dialogue that all too often goes 
 unrecorded, yet which reflects many of the concerns common to 
 all users and suppliers of information. Because of this, the partici- 
 pants agreed that the student-faculty dialogues would be re- 
 corded, with the caveat that wherever the material could be made 
 more meaningful, for the general reader, appropriate additions or 
 deletions would be made. 
 
 With nearly one-third of the Federal Government's statis- 
 tical dollars going to the Bureau, through direa appropriations or 
 reimbursable work for other agencies, it was only natural that 
 Bureau activities would play a central role in the discussion. Even 
 so, there were times when the speakers referenced the Census 
 
 Bureau only in passing, while at others the Bureau was the sole 
 subject of the presentation. However, all the discussions have 
 several points in common: 1) All the discussants have at one time 
 or another had the responsibility of providing information to 
 decisionmakers. 2) All the discussants have shared the frustration 
 of not having their information used to its fullest extent. 3) All 
 the discussants would like to see information used more ex- 
 tensively in the decisionmaking processes of our society. 4) All 
 the discussants have used the Census Bureau as a denominator as 
 they attempted to provide more meaningful measures to tell us 
 "where we are and wither we are tending." 
 
 As the title of this publication indicates, the Census Bureau 
 generates statistics that can serve as meaningful measures of 
 societal change. In "Aspects of the Census Bureau" Vincent 
 Barabba points out that these measures need not be complex and 
 can be as simple as basic mathematics. It is a fact that some of the 
 most meaningful statistical measures merely require the utiliza- 
 tion of both a numerator and a denominator. The numerator will 
 show the actual number of events that have occured, while the 
 denominator shows the population within which the events could 
 possibly occur. With this information analytical data can be 
 developed to compare a variety of relevant factors. Almost every- 
 one of the presentations addresses the problem of developing 
 meaningful measures, whether it is in the corporate, marketing, 
 health care, or national accounts fields. Today, the Census Bureau 
 is involved, through its advisory committees, public meetings, and 
 consultations with users, in attempting to understand and develop 
 the statistical tools to present society's decisionmakers with 
 appropriate measures. This publication is intended to reflect some 
 of the current problems and trends in this effort, both within and 
 outside the Bureau. 
 
 Because of the varied interests and expertise of the seminar 
 participants there is not a natural flow from one presentation to 
 another, just as the problems that the data providers and users 
 face lack a natural flow from one to the next. An index of special 
 ideas and topics is provided for those readers having specific 
 interests. Thus, they will be able to rapidly locate relevant sections 
 of the dialogues. 
 
ASPECTS OF THE CENSUS BUREAU 
 
 Vincent P. Barabba. Director, U.S. Bureau of the Census 
 
 "As I have already suggested, instead of collecting everything, inundating the 
 decisionmakers, and hoping that they can pick out the needed pieces, we should find 
 ways of transmitting information effectively in more limited and usable formats." 
 
 Vincent Barabba 
 
 Mr. Barabba: As we begin this entire seminar I would like to 
 present a brief theoretical discussion of the current and future 
 cost and utilization of certain types of information. I hope to 
 provide you with some criteria for judging the data systems, 
 techniques, and usages that will be presented to you during this 
 course. 
 
 To begin, it is self-evident that our country is entering a 
 squeeze regarding the "margin of error" that both the govern- 
 mental and private sectors will be allowed in conducting their 
 affairs. The energy shortage indicates how narrow and ever 
 decreasing this margin really is. We know that, viewed over time, 
 business, manufacturing, and governmental decisions can have an 
 irreversible ecological impact. Therefore, information that is 
 sufficiently accurate, timely, and easy to use must be available so 
 that the best decisions can be made to maximize the use of our 
 limited resources. 
 
 There are real costs in both time and money associated with 
 the collection of reliable and useful data, but from a societal 
 perspective the cost of statistical information is closely related to 
 the use. The relative cost to society of any data varies and is 
 dependent on the number of times it is used. This assumes, of 
 course, that we agree that society benefits when decisions relative 
 to the utilization of its limited resources are based on accurate 
 information. 
 
 The simple collection and publication of statistical data 
 does not by itself fully meet the needs of today's decisionmakers, 
 nor does the presentation of statistical data always maximize its 
 potential uses. Even though society's decisionmakers are fur- 
 nished with accurate, timely, complete, and usable data, it is a 
 disservice if they and society are surroundea by too much data. 
 Too much data places the user in the same plight as Coleridge's 
 Ancient Mariner, who cried, "Water, water, everywhere, nor any 
 drop to drink." It is necessary to find ways of reducing the 
 current oceans of data into pools of meaningful information for 
 decisionmaking. 
 
 To be truly useful, the data must be presented in a way 
 which will achieve meaningful measures. Meaningful measures do 
 not have to be complex. They can be as simple as basic 
 mathematics. Some of the most meaningful statistical measures 
 merely require the utilization of both a numerator and denomi- 
 nator. 
 
 The numerator shows the actual number of events that have 
 occurred— for instance the number of births, deaths, or divorces; 
 the number of whites, blacks, or persons of Spanish origin, the 
 number of families with incomes under $5,000 or over $15,000, 
 or other similar measures. 
 
 The denominator, on the other hand, will show the 
 population within which the events can occur. The denominator 
 may be the total population, the number of married women of 
 
 childbearing age, the total number of families, or some similar 
 population group. 
 
 For example, in health statistics it is important to know the 
 number of people who die from a specified cause, such as cancer. 
 The total number alone, the numerator, can provide useful 
 information on the magnitude of occurrence of this disease. 
 Additional understanding can be obtained by looking at the 
 denominator. The denominator could be the total population, the 
 total dying from all causes, or the total in one of the age, race, or 
 sex groups. Knowing the denominator permits the development 
 of analytical data showing the percent of the total population 
 dying from this cause, the percent these deaths are of all deaths, 
 the rates of occurrence for various population components and 
 the changes in these rates. 
 
 Systems' Costs 
 
 In order to understand the potential contribution of 
 statistical information and to compare the changing relative costs 
 of collecting differing types of data, let us examine our current 
 information systems by looking at three different ways we now 
 generate statistical information. The first is through specific- 
 purpose surveys that are designed especially for the problem to be 
 explored. The second is through the use of a more comprehensive 
 general-purpose survey or census. Rather than having a design 
 intended to produce information on one particular problem, this 
 type is constructed to provide insight into several areas of 
 concern. The third method is through the synthesis and analysis 
 of data already collected. The costs of these information system 
 components vary greatly. The costs will change over time, and 
 again, the cost per use will vary. 
 
 Figure 1 illustrates the change in relative costs for each use 
 of information acquired through the collection of statistical data 
 in a specific-purpose survey. The specific-purpose survey may 
 include in its universe both the numerator and the denominator 
 that are of primary interest to the decisionmaker. The costs of 
 collecting this type of data have steadily increased over the past 
 years and are likely to continue to rise in the foreseeable future. 
 The measurement of specific actions, attitudes, or individual 
 characteristics in relation to a particular problem or objective 
 requires sample survey design, questionnaire development, the 
 development and implementation of a collection methodology, 
 and the utilization of analytical techniques. 
 
 Specific-purpose surveys may also be directed at an already 
 identified population such as patients in hospitals, persons who 
 have previously been identified as recipients of pensions, or 
 similar unique universes. In some cases the population data 
 collected in a general-purpose census can then be used as the 
 denominator. 
 
 1 
 
ASPECTS OF THE CENSUS BUREAU 
 
 FIGURE 1. RELATIVE COST OF SPECIFIC 
 PURPOSE DATA 
 
 FIGURE 2. RELATIVE COST OF GENERAL 
 PURPOSE DATA 
 
 CHANGE IN 
 RELATIVE COSTS 
 "PER USE" 
 OF INFORMATION 
 
 YESTERDAY 
 
 TOMORROW 
 
 TIME 
 
 CHANGE IN 
 RELATIVE COSTS 
 "PER USE" 
 OF INFORMATION 
 
 YESTERDAY 
 
 — TOMORROW 
 
 TIME 
 
 Although extremely useful information can be obtained 
 through specific-purpose surveys, and the cost is held down by 
 sampling— they are often expensive undertakings. Survey costs will 
 continue to increase due to the rising cost of labor, transporta- 
 tion, the use of more sophisticated and expensive methodologies, 
 and the additional constraints that have been— and will be— placed 
 on data-collection activities by the public's concern for individual 
 privacy. One of the major reasons for the relatively high cost of 
 specially collected data is that the information obtained cannot 
 be used for other analysis and decisionmaking needs. A specific- 
 purpose survey has, by definition, a narrow range of uses and the 
 cost of generating the data is borne by a limited set of users. 
 Unless a data set can be used for a variety of purposes, the cost to 
 society remains the same as the cost and value to the original 
 user. 
 
 This next chart (figure 2) provides an illustration of the 
 relative "per use" cost of information collected in a general 
 purpose type of statistical system. Census data are a prime 
 example of general-purpose statistics collected for a multitude of 
 uses. General-purpose data can provide insight into a variety of 
 specific problems through the use of proper comparison and 
 extrapolation. General-purpose data provide benchmarks describ- 
 ing the general social or economic situation; they establish and 
 define parameters and provide a vital time reference. 
 
 Costs related to the collection and analysis of general 
 purpose data face the same pressures to increase that we find with 
 specific data. But, as the frequency of use increases, the cost per 
 each use is reduced. The collection of information in a decennial 
 census which requires contact with every household in the Nation 
 is, of course, an expensive undertaking. Census information is 
 widely used by Congress, Federal, State, and local governments, 
 businessmen, educators, and researchers. Many others are un- 
 aware that they use it, since census statistics are presented in a 
 pervasive abundance in textbooks, almanacs, magazines, and 
 newspapers. 
 
 These myriad uses make census statistics comparatively less 
 expensive on a "per use" basis. The use of self-enumeration, 
 whereby respondents receive and return the census questionnaire 
 by mail, provides for lower collection costs and more accurate 
 information. 
 
 Figure 2 tells us that if the ever increasing cost of collection 
 placed on society is not offset by an increased number of uses per 
 data item collected by society, the cost will continue to rise 
 (upper line). However, if the increased number of uses is equal to 
 the increase in cost, the cost remains equal to the per use cost of 
 yesterday (solid line). If the increase in uses is greater than the 
 increase in cost, the cost per use will fall below yesterday's cost 
 (bottom line). 
 
 The Census Bureau's Current Population Survey is some- 
 times perceived as a specific-purpose survey. Each month some 
 50,000 households are interviewed for the very specific purpose 
 of obtaining information on the magnitude of unemployment in 
 the Nation. If this was the only information collected, and its use 
 was limited to the statistics released by the Bureau of Labor 
 Statistics, it would obviously be a limited-purpose survey. 
 However, the CPS is used as a general-purpose survey to update, 
 on a national basis, information collected in the decennial census 
 on such matters as family income, educational attainment, school 
 enrollment, marital status, households and families, migration, 
 and similar subjects. 
 
 This ongoing survey also permits the addition of unique 
 questions on such subjects as birth expectations and voter 
 participation. Further analysis from annual average data as well as 
 special tabulations from CPS data can be made. 
 
 The use of administrative records, available from Govern- 
 ment and business, for the production of general-purpose, 
 anonymous statistics is one of the factors which helps reduce the 
 cost of data collection. The value of information from administra- 
 tive records is considerable because the expense of collecting the 
 information has already been allocated to other functions. These 
 records-whether used by public officials in their planning efforts 
 or by the business community— can be very valuable, particularly 
 on the local level, through the use of the Census Bureau's 
 geographic coding system, the GBF/DIME system. It is also the 
 case that the hardware and software expenses of the GBF/DIME 
 system have already been paid by research and development 
 funds. Again, the greater the usage the lower the cost of each use. 
 
 One of the major assets of the GBF/DIME system is that it 
 permits locally generated data to be related to census data, 
 opening unlimited opportunities for the user. The numerator, in 
 
ASPECTS OF THE CENSUS BUREAU 
 
 this case, is the administrative records and the denominator, the 
 census data. 
 
 However, I must point out that the effective use of masses 
 of information involves certain drawbacks. Noteworthy is the 
 question of confidentiality-an issue of great concern to Congress, 
 the media, and the public. The Census Bureau has a legal 
 obligation to protect the individual's right to privacy. This is 
 based on Federal statute and long standing Bureau principle. The 
 GBF/DIME system cannot provide confidentiality at the local 
 level— only the actions and intent of those responsible can do 
 that. However, the GBF/DIME system does allow specific 
 characteristics to be assigned to a geographic area without 
 revealing personal information about specific people who live 
 there. The system has the merit of permitting any number of 
 applications; from Federal to local and private areas there are uses 
 which can be of immeasurable benefit to society. 
 
 For example, the GBF/DIME system permits a community 
 to take its local files such as those relating to births, deaths, 
 morbidity and health services— the numerators— and geocode 
 them to census tracts or smaller areas, if necessary, where there is 
 a population-at-risk. This can produce a wealth of information 
 about every small area in the community, since the file contains 
 the number of deaths and their causes, births, the number of 
 health services, and the people who receive the services. 
 
 Once these local data are structured through an information 
 system, they are useful in identifying populations which are most 
 vulnerable to health problems. The health information system can 
 also identify the specific needs of these people and recipients 
 of services. 
 
 A very important factor in this system is that if these 
 identifications are noted periodically, it is possible to monitor 
 trends to determine if needs are being met, services delivered, or 
 if other indications of quality health care are present. This ability 
 to monitor health trends is particularly valuable in predicting the 
 probable health conditions of those segments of the population 
 most prone to poor health. 
 
 This chart (figure 3) illustrates the likelihood that the 
 relative cost of analyzing data already collected may decrease in 
 the future. One means of making greater analytical use of data 
 already collected is through the increased application of com- 
 puter capabilities. Computer capability has kept ahead of the 
 increased costs of purchasing or renting comouter systems. In 
 addition, more qualified personnel are being trained to use this 
 advanced equipment. While the generalized software packages 
 have increased efficiency, access through remote terminals per- 
 mits more people to use expensive main frames. 
 
 The further analysis and synthesis of information already 
 collected also help provide some cost reduction by eliminating 
 false starts and needless repetitions. 
 
 A tremendous amount of statistical information has been 
 collected and published, and its further analysis can provide 
 valuable information to serve the social and economic needs of 
 our Nation. New mathematical techniques, such as model 
 building, simulation, synthetic estimates, and path analysis can 
 help provide more meaningful measures in the further analysis of 
 previously collected material. Moreover, the further analysis and 
 refinement of data, even when new mathematical techniques are 
 not used, can still provide additional understanding. The presenta- 
 tion of newly computed rates and ratios and the selection of the 
 most appropriate numerator and denominator can also help 
 provide meaningful measures. 
 
 FIGURE 3. RELATIVE COST OF ANALYSIS 
 OF AVAILABLE DATA 
 
 CHANGE IN 
 RELATIVE COSTS 
 "PER USE" OF 
 INFORMATION 
 
 YESTERDAY 
 
 TIME 
 
 — TOMORROW 
 
 The message of this next chart (figure 4) is quite straight- 
 forward. It summarizes the three previous charts and clearly 
 indicates that relative costs must be carefully considered. At some 
 point in time, when the combined costs of general-purpose data 
 and the further analysis of that data become sufficiently less than 
 the cost of specific-purpose data, the decisionmaker may very 
 well decide to use the less costly approach. He will, however, still 
 need to give serious consideration to the quality, availability, 1 
 cost, timeliness, and respondent burden in determining the type 
 of data to select. 
 
 Even though this chart may indicate some cost-benefit gains 
 through the use of other types of data, specific-purpose surveys 
 will continue to be conducted, for there will always be the need 
 for unique information. This chart demonstrates that there will 
 be several alternatives in the selection of information needed for 
 decisionmaking processes. 
 
 FIGURE 4. RELATIVE COST OF DATA 
 
 CHANGE IN 
 RELATIVE COSTS 
 "PER USE" OF 
 INFORMATION 
 
 YESTERDAY 
 
 — TOMORROW 
 
 TIME 
 
 Administrative records may be available and valuable as a resource, yet 
 privacy and confidentiality rules or considerations may prohibit access. 
 
ASPECTS OF THE CENSUS BUREAU 
 
 The next chart (figure 5) illustrates the relationship 
 between the degree of certainty a decisionmaker requires before 
 he is willing to make a decision and how much that certainty will 
 cost in time and money. The vertical axis measures the degree to 
 which a particular phenomenon, circumstance, or experience can 
 be explained. The horizontal axis measures the number of 
 variables necessary to reach a particular degree of explanation. 
 
 FIGURE 5. DATA FOR THE DECISION MAKER 
 
 THE 1970S THE COMPUTER 
 AND THE DEVELOPMENT OF MORE 
 MEANINGFUL MEASURES 
 
 DEGREE OF 
 EXPLANATION 
 OF A 
 
 PARTICULAR 
 PHENOMENON 
 
 THE 1960S... 
 THE COMPUTER 
 AND "MORE" DATA 
 APPROACH 
 
 NUMBER OF VARIABLES 
 
 (TIME AND COST OF DECISION MAKING) 
 
 During the 1960's most decisionmakers found their solu- 
 tions in the enormous amount of data generated by the computer. 
 This approach is typified by the lower curve, illustrating that 
 many variables are required before a large degree of a particular 
 phenomenon can be explained. This reflects the approach of "give 
 me everything you can get on the subject." 
 
 However, one of the problems with this approach was the 
 difficulty in using the reams of data that poured out of these very 
 efficient data generators. This difficulty led to increased efforts 
 to develop methodologies which would allow decisionmakers to 
 deal with only the more relevant and meaningful variables. This 
 latter approach is reflected in the upper curve, which illustrates 
 that certain variables, properly calculated, can contribute to a 
 greater degree of explanation. In a sense, the upper curve reflects 
 an effort to turn the masses of data illustrated by the lower curve 
 into information for decisionmaking. The development of eco- 
 nomic indicators is an example of this. 
 
 In summary, there will continue to be specific-purpose 
 surveys, but greater cost efficiency and public concern over 
 privacy will cause the further growth and development of 
 large-scale general-purpose surveys and the continued refinement 
 and analysis of data already collected. The growing statistical 
 needs of the Nation, such as the social and economic status of the 
 population, and measures of our economic activity and natural 
 resources, clearly indicate the vital need for the continued 
 development of statistical information systems to meet the varied 
 needs of both national and local decisionmakers. 
 
 Student: It seems one of the Census Bureau's major problems is 
 the dissemination of the data collected. Is the cost of distributing 
 this information prohibitive? 
 
 Mr. Barabba: There are cost problems, but they are not the major 
 concerns. The problem is the disservice to decisionmakers by 
 
 inundating them with too many facts. At best they've got to go 
 through a whole roomful of facts to find the two or three they 
 really need. If the providers of information have done a good job, 
 they will have assessed and transmitted only the information 
 needed for decisionmaking. But, that's a lot easier to say than to 
 do, since most people don't know what they need, and it is often 
 difficult to determine who users and potential users may be. 
 
 Student: By controlling or limiting what information is collected, 
 you're making qualitative or quantitative judgments. These 
 judgments can influence very directly the kind of decision that's 
 going to be made. Is that a problem? 
 
 Mr. Barabba: It's a considerable problem, based on who's your 
 most important constituent? Are the professors who are doing 
 theoretical work, or local planners who may want specialized 
 information emphasizing local data the Bureau's most important 
 constituency? It's difficult to serve both users equally, be- 
 cause they want different kinds of information. 
 
 The professor is generally looking at national concerns 
 because he's formulating theories. However, the city planner in 
 Wichita Falls asks, "Why don't you take the census every other 
 year, because I need detailed information about the blocks in my 
 town?" With our limited budget we cannot satisfy both needs 
 with ideal and exact service. So we attempt to compromise, and 
 that's where the difficult part lies. 
 
 The Bureau spends, of its appropriated money and the 
 work done for other people, over $130 million a year. That's a lot 
 of money and a lot of information. But, we really don't have a 
 good criterion to evaluate whether we're doing the best job for 
 society, because society has not told us through its elected 
 representatives what it really needs. In other words, it seems that 
 the only time we learn what's really needed is when people 
 complain that certain data don't exist. We use our advisory 
 committees to a great extent, and they tend to cover the broad 
 spectrum of information users. Then, too, Congress does a pretty 
 good job by holding open hearings and asking us why we spend 
 the money the way we do. Each Administration has certain things 
 that it would like to accomplish. For example, the New 
 Federalism has done more to create the need for local area data, 
 and my ability to argue for more of it, than anything else. This is 
 mainly because we've been able to make a case to the Office of 
 Management and Budget that, besides giving these people their 
 share of the dollar, the Federal government needs to provide the 
 statistical tools to spend it wisely via an assessment of the 
 program's impact. 
 
 For example, last year we had a small pilot program in San 
 Francisco, Region 9, where we had two Census employees 
 attached to the Federal Regional Council. Any time the council 
 had an information problem these employees were available to 
 bring together census and other data for them. They were there 
 to help put together the information packages for various policy 
 decisions. 
 
 This is a step in the right direction, because we are 
 beginning to distribute specific information to those places where 
 the decisions are being made. And we're not simply providing the 
 national average income figures or the national crime rate, nor are 
 we forcing officials in San Francisco or Boston to spend their 
 dollars on the basis of these national estimates. To provide data 
 sets to these various areas is a lot easier said than done. The raw 
 data is available; the problem is providing the information in a 
 retrievable format and understandable language. 
 
ASPECTS OF THE CENSUS BUREAU 
 
 As I have already suggested, instead of collecting every- 
 thing, inundating decisionmakers, and hoping that they can pick 
 out the needed pieces, we should find ways of transmitting 
 information effectively in more limited and usable formats. 
 
 Student: You have implied that the types of data the Census 
 Bureau produces have significance at several levels of use and, 
 obviously, many of these potential uses have political implica- 
 tions. Unemployment would be a graphic example of that. Do 
 you ever find instances where there's pressure put on you to 
 either change the parameters of your criteria for categorization or 
 to come up with some statistical results to validate a prejudice or 
 a viewpoint, either from the executive or the congressional side? 
 
 Mr. Barabba: Never on the actual data. 
 
 Student: Interpretation, perhaps? 
 
 Mr. Barabba: First of all, and you'll have to take my word for 
 this, any attempt to change actual numbers is virtually impos- 
 sible. The reason it's impossible is that we do the collection and 
 processing of the unemployment statistics in 15 days. Then, it 
 would take collusion on the part of probably 120 civil service 
 workers all the way from the supervisor in the field, through the 
 regional director, and then through all the mathematical statisti- 
 cians in the Census Bureau who deal with this particular survey. 
 These figures are checked at various points aiong the way. Even if 
 we could pull it off inside the Census Bureau, the figures have to 
 go to the Bureau of Labor Statistics, which is an organization 
 independent of the Census Bureau, and they have their own 
 analytic techniques. 
 
 Student: You're saying that within your organization there isn't 
 a problem. But how about externally, when somebody takes the 
 data or suppresses its publication? 
 
 Mr. Barabba: There's a circular called OMB A-91 which describes 
 the procedures for information release. Also, the Federal Register 
 publishes a schedule of reports and the release dates. Finally, no 
 Administration spokesman can speak about those statistics until 
 an hour after they have been released to the press. 
 
 Student: Who decides, for instance, on the statistical description 
 of poverty? One administration could possibly think that there's 
 too much poverty in the United States and might decide to raise 
 the figure on the poverty level. Who makes the ultimate decisions 
 concerning what is measured, and the levels that represent 
 "poverty" or other subjective states of being? 
 
 Mr. Barabba: Basically, the poverty line is a floating number tied 
 to the Consumer Price Index (CPI). Poverty was benchmarked at 
 a point in time and is now updated on the basis of what the 
 Consumer Price Index is. 
 
 Student: I'm not as interested in the number or how it's used as 
 in how the original benchmark is arrived at. 
 
 Mr. Barabba: Poverty statistics are based on a definition de- 
 veloped by the Social Security Administration in 1964 and 
 revised by a Federal interagency committee in 1969. 
 
 Student: Does Congress have any input? 
 
 Mr. Barabba: You bet. Hearings. And legislation. If they are 
 dissatisfied with a procedure, they can legislate how it should be 
 done. If they can't get it that way, they always have something 
 called the "budget." The budget is why departments and bureaus 
 tend to be rather responsive to the Congress. The Congress can 
 really affect your operations by delaying the signing of your 
 budget, forcing you to work on a "continuing resolution" that 
 basically says that you can't do anything that you didn't do 
 during your last approved budget. 
 
 Student: How much time does Congress devote to supervising 
 the Bureau? Do they depend on you for that information? 
 
 Mr. Barabba: Within the Post Office and Civil Service Committee 
 of the House of Representatives there is a Census Oversight 
 Subcommittee. It has a chairman and eight members. In the 
 Senate it's the Post Office and Civil Service Committee, but it 
 dosen't have a specific subcommittee looking over Census 
 operations as such. 
 
 Sampling 
 
 Student: Who decides which households to survey and how is 
 that decision arrived at? 
 
 Mr. Barabba: First, we use a probability sample. That means, 
 when we start out, everybody in the defined universe has a 
 known opportunity of being drawn into the sample. You have 
 Mr. Morris Slonim's book on sampling. I don't want to get into a 
 detailed discussion on probability sampling since you can find the 
 treatment of this subject in a number of excellent texts. Let it 
 suffice to say that the sample is randomly drawn, and appropriate 
 field procedures are devised so that human intervention is 
 reduced to a minimum, resulting in an estimate which has the 
 least possible bias and sampling error consistent with the available 
 resources. 
 
 Facetiously, if you don't believe in sampling, the next time 
 you have your blood tested, have them take all of it. Because 
 that's really what you're doing. You're taking a sample of 
 something and you're assuming that it represents the total 
 universe. 
 
 Now, how do we go about selecting a sample that will not 
 be biased? The strength of the probability sample lies in the fact 
 that we have no established quotas as to whom we will interview. 
 That is, we let the laws of probability determine who will be 
 interviewed rather than allow the interviewer to use his or her 
 judgment in obtaining a specified number of interviews from 
 designated subpopulations. Conceptually, the simplest way of 
 doing this would be to have a list of everybody in the United 
 States and give each of them a number; then draw the sample size 
 desired from a list of random numbers, which is a list where the 
 numbers have no relationship to each other. If information were 
 collected properly from members of this sample you would be 
 able to draw inferences about the universe of all people in the 
 United States with a specified confidence. Since we will not be 
 able to cover the elements of a course in probability sampling 
 here, you must accept some of these principles on faith, unless 
 you want to actually go out and select a sample. We can arrange 
 that, too. 
 
ASPECTS OF THE CENSUS BUREAU 
 
 Student: If it's anything like precinct work, no. 
 
 Mr. Barabba: Precinct work, for a political party, would be an 
 example of the worst kind of sampling for drawing inferences 
 about the general population. The person you send out is a very 
 partisan person who is going to people who generally hold similar 
 beliefs. And, if the people don't, the worker is supposed to try to 
 persuade them. But, we want to find people representative of the 
 whole community and, for purposes of sampling, we don't care 
 where they live; that's why we select them randomly. If we had 
 to interview 10 people from a particular sampling unit, and we 
 went right down the block and took them in sequence, we would 
 have a biased sample. For example, all 10 might be found in one 
 high-rise apartment; thus, no one else, for example those in single- 
 dwelling units, would have a chance to come into the sample. 
 
 Here's another example of the type of problem that could 
 arise. A survey was conducted in England and after they 
 randomly selected street intersections they interviewed only 
 those houses located on corner lots. Consequently, they ended 
 with a sample of people who lived on the corners of the blocks. 
 Now, in England as in the United States, the corner lot is usually 
 the most expensive lot, so they had built an automatic bias into 
 their sampling design. So, we have random procedures which 
 force enumerators to interview specified households. Once they 
 get to the door that's identified for them, their interview sheet 
 specifies the person in the household to be interviewed. Depend- 
 ing upon the purpose of the study, this could be the oldest 
 person, or the youngest, or one selected at random. There's a set 
 of procedures for doing that as well, so that you don't have the 
 interviewer talking only to people with whom she or he feels 
 most comfortable. 
 
 Student: Well, where do you get the names-telephone books, 
 voter registration, or people who have houses? 
 
 Mr. Barabba: Let me tell you how the Census Bureau does it. Let 
 me use as an example the Current Population Survey, which is 
 our largest and most comprehensive continuing survey. 
 
 If we had a complete and up-to-date list of addresses of all 
 housing units in the United States, the problem of designating a 
 sample of housing units could be approached as follows. We could 
 count through the list and designate every 14 hundredth housing 
 unit for the sample. If we counted through the entire list of 
 roughly 70 million housing units, this process would designate a 
 sample of about 50,000 units, very nearly the size of the current 
 CPS. 
 
 The preceding, although very much oversimplified, illus- 
 trates the concept used at the Census Bureau to define the CPS 
 household sample. As described, however, such a plan is not 
 really economically feasible. In the first place, a sample selected 
 in this way would require a visit to at least one sample unit in 
 almost all of the 3,141 counties of the country and would require 
 expensive field efforts for data collection. Secondly, the list of 
 housing units for all areas of the United States would have to be 
 developed and maintained, a very costly operation. However, if 
 these efforts could be confined to a sample of primary areas, and 
 if such a list were constructed within a subsample of smaller areas 
 chosen out of these primary areas, the system becomes economi- 
 cally and practically feasible. This, in fact, is the system used by 
 the Census Bureau. 
 
 To select the sample of primary areas, the entire United 
 
 States is divided into primary areas called primary sampling units 
 (PSU's). PSU's are designed for use as convenient interviewer 
 assignment areas and at the same time to be efficient sampling 
 areas for statistical purposes. We define a PSU as either an SMSA, 
 a county, or a group of counties. A total of 1,924 PSU's have been 
 designated in the CPS of the 50 States and the District of 
 Columbia. 
 
 The second step involves setting up groups of PSU's called 
 strata that are homogeneous with respect to the statistics we are 
 primarily interested in collecting. In the case of the CPS, the 
 grouping attempts to make PSU's within the strata homogeneous 
 with respect to labor force data. The largest 156 PSU's (mostly 
 SMSA's) were each made into a stratum; nearly 65 percent of the 
 U.S. population resides in these strata. The remaining PSU's were 
 grouped into 220 strata, each of which contained a number of 
 PSU's whose population totaled about 350,000. 
 
 The sample of primary areas is then defined as all of the 
 156 largest PSU's plus a sample of one PSU from each of the 220 
 remaining strata. The PSU's are drawn from these latter strata 
 with probabilities proportionate to their size, the larger the 
 proportion of the stratum population contained in a PSU, the 
 larger its probability of selection. An additional sample of 110 
 PSU's was drawn from half of the 220 strata, making a total of 
 330 PSU's selected from the 220 strata. Because it is possible for 
 a PSU originally selected out of one of the 220 strata to appear 
 again in the second sample, there are a total of 461 different 
 PSU's in the sample (including the 156 large PSU's). These sample 
 PSU's comprise a total of 923 counties (about 76 percent of the 
 total U.S. population). 
 
 Within each of these sample PSU's, a sample of smaller 
 areas, called enumeration districts (ED's) is selected, and it is 
 within these ED's that the current register of housing units is 
 needed. The register is made up from three types of lists: Units 
 enumerated in the 1970 Decennial Census, units missed in the 
 1970 Decennial Census, and units constructed since the 1970 
 census. A record of units enumerated in the last census is 
 available for sampling in all areas of the country including the 
 sample of ED's in the 461 areas designated above. 
 
 Units constructed since the census are measured by two 
 methods. For most of the United States (covering about 87 
 percent of the construction), a building permit is required by a 
 local civil jurisdiction, and the Census Bureau visits a sample of 
 these jurisdictions to select a sample of the new units represented 
 by the permits issued. For the balance of the United States, new 
 construction is measured by a sample of small land area segments 
 within which the existing units are canvassed to identify those 
 built since the census. For some ED's, where records of units 
 enumerated in the census have incomplete address information 
 (incomplete street name and house or apartment number), the 
 sample of land areas is used also to represent units enumerated in 
 the census rather than the sample drawn from the incomplete list. 
 
 Units missed in the 1970 census are represented by a 
 sample of small land areas. Just after the census, an intensive 
 canvass of units in these segments and a match against units 
 recorded in the census were made to identify units which should 
 have been enumerated in the census but were not. 
 
 The lists are used to designate a random sample of units for 
 interview by a set of rigid mechanical steps which do not allow 
 subjective judgment, personal choice, or convenience of obtaining 
 an interview to enter the selection process. The procedure 
 maintains the concept of counting through all units of the United 
 
ASPECTS OF THE CENSUS BUREAU 
 
 States by varying the counting intervals to insure a constant 
 probability of selection for each unit. 
 
 You have two main categories of error from surveys. They 
 are usually designated merely as sampling and nonsampling 
 errors. Sampling error arises because the estimate from an 
 individual sample is not the same as the average of all possible 
 samples. If the sample is drawn with known probability, and if the 
 survey is carried out properly, there is a well-known body of 
 theory which permits you to estimate this sampling error. Given 
 the sampling error, it is possible to make inferences about the 
 estimates with specified confidence. For example, two sampling 
 errors above and below the estimate give an approximate 
 95-percent confidence interval. What this means is that if 100 
 samples were taken and 95-percent confidence limits constructed 
 on each sample, approximately 95 of them would include the 
 average (but unknown) value of all possible samples and 5 of 
 them would not include this average value. 
 
 Nonsampling error, which includes bias, is much more 
 difficult to measure, and the theory for estimating the compo- 
 nents of nonsampling error is not as well developed as it is for 
 estimating sampling errors. Nonsampling errors can never be 
 eliminated. They occur in the way a question is asked, the time of 
 day the interview occurs, and the characteristics of the respond- 
 ent and the interviewer. In some cases you can get a measure of 
 the influence of age of interviewer, for example, by assigning 
 respondents at random to interviewers of different ages and 
 estimating the difference in response by age class. If you are 
 talking about variables such as income, you must go to more 
 elaborate procedures like auditing a random subsample of 
 respondents, provided you can get them to agree to an audit. 
 
 mented, on a pilot basis, with paying people to fill out the diary 
 and there was no appreciable difference in the response rate. 
 However, commercial interview firms have used that approach 
 successfully. 
 
 What you have going for you in the polling and survey 
 business, in the Government and nongovernment research as well, 
 is that people are not normally given the opportunity to 
 express their opinions. I can recall that my wife never believed 
 that I had a real job until she was interviewed. She'd always said 
 nobody would ever answer our questions, and yet she found 
 herself answering, mainly because it was interesting, and no one 
 had ever asked her opinion before. 
 
 Student: Does the reliance upon mailed-in questionnaires rather 
 than personal interviews mitigate against obtaining responses? 
 
 Mr. Barabba: In the census, about 87 percent of the people who 
 received their questionnaires by mail sent them right back. That 
 figure says that we only had followup on 13 percent of that 
 population. We were able to take our resources and, in a sense, 
 segment the population: Those who were easy to get the 
 information from and those who were more difficult. Then we 
 focused on the balance. 
 
 One of the crucial aspects of the door-to-door technique is 
 the enumerator could be somebody from your neighborhood, and 
 you might be more hesitant to answer the questions. Through the 
 mail-out method, you can fill out your questionnaire in the 
 privacy of your home, put it in a confidential envelope, and mail 
 it back to the Census Bureau. There are a lot of advantages to the 
 mail-out/mail-back format. 
 
 Regarding General Response 
 
 Student: Are you having any trouble getting people to respond 
 to Bureau surveys? I have read that private polling organizations 
 are encountering more and more resistance. 
 
 Mr. Barabba: Our Current Population Survey is a good indicator 
 of whether we have a problem, because in this survey we 
 interview some 50,000 households every month. It's a voluntary 
 survey. Less than 2 percent of the people contacted refuse to 
 participate in the survey. That figure has remained fairly stable 
 through all the political problems of the recent years. It has cost 
 us a little bit more time at the door, two or three more callbacks, 
 and more intensive followup procedures to maintain that level. 
 
 Student: Are there any other surveys demanding public response 
 besides the decennial census? 
 
 Student: Do you have a followup to cross-check the mail-back 
 responses? 
 
 Mr. Barabba: There's a whole series of evaluations on the 1970 
 census which are in the process of being published. Many of these 
 evaluations cover the extent to which different items bias the 
 response. I think Mr. Levine could deal with those more 
 effectively. 
 
 Student: Has anyone actually been prosecuted for giving false 
 information? 
 
 Mr. Barabba: Not false information, no. There have been very few 
 prosecutions; but when there is, it is someone who will not fill 
 out the form. 
 
 Student: Oh, you really will prosecute? 
 
 Mr. Barabba: Any Bureau survey conducted more than once a 
 year is voluntary. All our economic census activity and the census 
 of agriculture are mandated surveys, the responses to which are 
 mandatory. 
 
 Student: Do you ever pay anybody or offer them a free gift to 
 respond? 
 
 Mr. Barabba: Oh, yes, we have to. If we let nonresponse become a 
 way of life, there could be a real problem. However, we normally 
 don't go after a person until he says, "Now, look here, my fellow 
 Americans, I told those Bureau snoopers to go jump in a lake and 
 you should tell them to do the same thing." We have to deal with 
 that. If our data are not accurate the whole base of the entire 
 Federal statistical information system is also inaccurate. 
 
 Mr. Barabba: Oh, no; although, as a matter of fact, there was a 
 study called the Consumer Expenditure Survey in which one 
 filled out a diary and was then interviewed quarterly. It took 
 about 2 hours every time you were interviewed. We experi- 
 
 Student: What do you do about mobile people? 
 
 Mr. Barabba: We have a couple of problems when it comes to 
 people who aren't counted that are quite distinct. They arise 
 
ASPECTS OF THE CENSUS BUREAU 
 
 when people are not counted because they are not available, or 
 because they don't want to be counted. 
 
 Now, when they are not available, the Bureau will 
 continually attempt to contact a household until somebody 
 answers the door. Let's say somebody's on a 3-month vacation. 
 After so many times of knocking on the door, we'll go to the 
 next door neighbor and ask how many people live in that house. 
 Then we will record only that information, plus a few basic 
 characteristics. We'll also determine whether it's a single-family 
 dwelling unit. 
 
 If we can't get answers to the question, we'll impute 
 information for that unit by replicating the data for an 
 enumerated unit. We substitute the characteristics for the 
 previously processed unit. Understand that imputation amounts 
 to an infinitesimal number so that the impact on the total census 
 is minimal. 
 
 Minority Undercount 
 
 Student: For specific groups, I understand that your count was 
 pretty far off. For the lower income groups there's been a lot of 
 publicity that your procedures are severely biased. 
 
 Mr. Barabba: I'll give you my very famous speech on the 
 undercount. First of all, I should point out that the U.S. Census 
 Bureau— not that we're defensive or anything— not only admits 
 that it made a mistake, but quite precisely estimates the extent to 
 which it erred. That undercount number that everybody talks 
 about was provided by the U.S. Census Bureau for evaluative 
 purposes. 
 
 There are several ways of estimating the undercount, and 
 you have to remember it's an estimate. However, we feel, given 
 the mail-out/mail-back approach to the census in urban areas, 
 that we hit almost every household. But, if there are five people 
 living in a household— really living in a household— and we only 
 counted four, that's very difficult for us to deal with, particularly 
 if the fifth person doesn't want to be counted. 
 
 Those of you who are involved in welfare activities realize 
 that it is in some individuals' vested interests not to be identified 
 with a particular household, as otherwise their welfare payments 
 would be cut off. Those of you in law enforcement realize that 
 for an illegal alien it is not in his best interest to be identified at 
 all. These types of individuals have not had an opportunity to 
 read "Privacy and the Need to Know," and therefore they do not 
 understand that such information will not get in the hands of any 
 other agency of Government. 
 
 There is another kind of disclosure problem that also may 
 affect people. Let's say there is a rule in a public housing unit 
 that you can have no more than six people per apartment. We 
 came along and, in counting the apartment occupants, the 
 numbers that we record may show that some apartments are in 
 excess of the rule. Let's say this apartment complex was an entire 
 block. We do generate block statistics. The manager could go 
 down the block statistics and determine that there was an excess 
 of people living in his apartments. At that point we have released 
 information that some would argue is of detriment to those 
 individuals who are violating the rules. We have not given away 
 which apartments they were. Nobody can get their hands on that. 
 And we may not have known that there even was a rule. But we 
 certainly have made the violators more vulnerable to discovery. 
 
 Therefore, one of the precautions we take to lessen the 
 chances of this sort of thing happening is to have a disclosure 
 
 rule. The rule is that if there are so few people living in an area 
 that you could, from inference, determine the characteristics of 
 an individual or housing unit, we will suppress that number or 
 limit it to being released in a much larger context which makes it 
 virtually impossible to identify anyone. 
 
 Remember, there is no way that the U.S. Census Bureau, 
 without abrogating all the traditions of our society, can force a 
 person to tell us how many people are living there, if for no other 
 reason than it would affect the reputation we have with all these 
 other people. 
 
 Also, remember that I told you that we prepare an estimate 
 of the population for evaluative purposes. In 1970 we believe that 
 the undercount amounted to about 2.5 percent of the total 
 population. We develop the undercount estimates for various 
 population groups (i.e., by age, race, and sex) by taking all the 
 available independent sources of information such as births, 
 deaths, immigration, emigration, Medicare statistics, and the like, 
 and estimating what the population ought to be. 
 
 You may say, "Why, if you can do that, do you bother to 
 take the census at all?" The answer is simple. We can only tell 
 you about the United States as a whole, and we can't tell you 
 about Orange County, Calif. That is why we have to take the 
 census. Besides, I can't tell you what the income is and can't tell 
 you all those other important things. 
 
 The Census Bureau has to educate the public that they will 
 not be hurt by giving us correct numbers. In fact, they are hurt 
 more if they don't give us the numbers, because these people are 
 not represented in revenue sharing calculations. They're elimi- 
 nated from the concept of equal elected representation. That is 
 the constitutional reason we take the census— to apportion the 
 House of Representatives equally and fairly among all people. 
 
 To help deal with undercount problems we are establishing 
 two committees. While the undercount is 2.5 percent among all 
 people, nationwide among black people the estimated undercount 
 is 7.7 percent. We don't know what it is among Spanish-origin 
 people because there are no current vital statistics on the 
 Spanish-origin population. It is not available on Spanish origin, 
 mainiy because it's very difficult to define "Spanish origin." We 
 are setting up a Black and a Spanish-origin advisory committee, 
 representative of local neighborhoods and various organizations 
 with Spanish and Black interests. The leaders of these communi- 
 ties have come to understand the value of these census numbers. 
 They know what they mean in dollars, and they know what they 
 mean in political representation. We want to, in a sense, develop a 
 communications system, so that the committee can go and talk to 
 their people and convince them that they all ought to be counted. 
 If that happens we'll have a better count. 
 
 You're right, there is an undercount, although it's not 
 completely a methodological problem. It is certain people's 
 unwillingness to be counted, for various reasons. We have the 
 authority from the Congress to enforce a count, but our 
 judgment is that if we ever attempted it, it would probably have a 
 great negative impact on the society. 
 
 Individuals, Files, and IRS 
 
 Student: Do you have access to individual files? If somebody 
 asked what a man had stated as his income on his census form, 
 could you go to the files and check? 
 
 Mr. Barabba: I have access to the individual returns, but my legal 
 authority allows me that access only insofar as I am carrying out 
 
 8 
 
ASPECTS OF THE CENSUS BUREAU 
 
 the Bureau's mission. Even so, access to the returns is severely 
 circumscribed by the very nature of our organization. Some of 
 the same problems exist here as would if anyone were to attempt 
 to change the unemployment figure. 
 
 For the 1970 census we have microfilmed copies located at 
 Jeffersonville, Ind., and Pittsburg, Kans. Normally, we would 
 have one chance in five of having detailed personal information. 
 Also, I must know where the person lived in 1970 in order to 
 search our records, because the questionnaires are not kept 
 in alphabetical order, but are indexed by county, city, and 
 street. We could call Jeffersonville and locate the microfilmed 
 questionnaire by checking the address register. The central 
 safeguard of this operation is that only Census Bureau employees 
 handle the material, and the penalty is a $1 ,000 fine or 2 years in 
 jail , or both, for unlawful disclosure. While even more important is 
 the effect such a disclosure would have on the Director's and the 
 Bureau's reputation. Also, I am the only political appointee out 
 of some 6,000 employees, so there is very little likelihood of my 
 commiting such an act without anyone knowing about it. So, 
 would you really want to come to me for that sort of 
 information? 
 
 Student: You wouldn't be a good source but, does the Bureau ask 
 for the Social Security number in the census? 
 
 Mr. Barabba: No, not on the census, which is mandatory. We do, 
 however, ask it in the Current Population Survey, which is 
 voluntary. 
 
 Let me ask you the question, why should we have it? 
 Besides the negative aspects of the invasion of privacy, what are 
 the positive aspects of the Bureau having access to the Social 
 Security numbers of our population? 
 
 Student: You would know where they were issued and could see 
 any geographic movement among those numbers. 
 
 Mr. Barabba: That would be one possibility. You'd have some 
 idea of when it was issued, too, wouldn't you? 
 
 We talked about bias earlier and said we could measure 
 sampling error, but it's very difficult to assess respondent bias. 
 One question we are frequently asked is whether people tell the 
 Census Bureau something different than they tell the Internal 
 Revenue. It's important to note the Census Bureau doesn't want 
 an answer about each individual, nor do we need it. 
 
 We could evaluate the accuracy of what people are telling 
 us about their incomes by matching our records with those of the 
 IRS. It may be that we would discover that the difference 
 between Census and IRS income statistics is the result of 
 different definitions. 
 
 Student: Isn't the real issue whether the Census Bureau should, in 
 fact, be used as an agency of "law enforcement" to turn over 
 certain information to the IRS for prosecution? Not necessarily 
 individual information, but even if you're saying that you only do 
 it in terms of groups. Is the matching of Census and IRS records 
 an appropriate means to solve problems of comparative defini- 
 tions? Aren't there other agencies who should more appropriately 
 be involved? 
 
 Mr. Barabba: There are. But who should have the final say on 
 what procedure we ought to use? 
 
 Student: It's not a question of procedure, it's a question of 
 powers. Should the Census Bureau be given the powers to assist 
 regulatory agencies? 
 
 Mr. Barabba: Who should determine whether the Census Bureau 
 should be given those powers? 
 
 Student: The people through their elected officials. 
 
 Mr. Barabba: If we determine there's a better way of collecting 
 data, then you are saying, as a citizen, you ought to be able to 
 modify our procedures by operating through the Congress. That's 
 exactly how the system operates. The Census Bureau is mandated 
 under authority of Title 13 of the U.S. Code. Our powers are 
 quite explicitly stated as to what we can and cannot do and our 
 authority to determine what is mandatory and what's voluntary. 
 Any time a Congressman feels that we have gone beyond our 
 authority he has the right to have us explain what we are doing. 
 
 For example, because so many Federal programs are based 
 on median income levels for an area, it might be advantageous to 
 determine the degree of difference between Bureau and IRS 
 income statistics. The Bureau would be willing to share its 
 information, stripped of all individual identification. This coop- 
 eration should not be misunderstood as assisting the IRS in law 
 enforcement, because we are protecting the confidentiality of our 
 respondents. In fact, the information given to the IRS would 
 consist of the socioeconomic characteristics of groups of people. 
 
 The study might indicate the tax system is not equitably 
 collecting money from all segments of society. Or, it may reveal a 
 social or educational problem that causes people to incorrectly 
 fill out the Bureau or IRS forms. 
 
 It is my choice to do it— and take the heat from some 
 Congressmen if it's a bad decision— or not— and take the heat 
 from other Congressmen who say they need the information. 
 That authority is well within the Bureau's rules and background 
 experience. 
 
 Ironically, we are a great taker of records from other 
 agencies. The economic census represents some 5.5 million firms 
 in the United States. But only 2 million of these were actually 
 interviewed. The other 3.5 million were small businesses, and we 
 got all the data on them from the Social Security Administration 
 and the IRS. 
 
 As long as we ascribe the same level of confidentiality that 
 is ascribed to the information by the IRS it seems to me that we 
 save money in the collection of information, and we save the 
 respondent the burden of filling out the form again. 
 
 But what do we give back? As I said, nothing in the way of 
 individual information, only socioeconomic characteristics of 
 groups of people. Also, we report that census data tends to be 
 overstated or understated from IRS data by such a factor. We're 
 going to have a legitimate piece of information that planners 
 ought to have. 
 
 We actually did something like this last year. The 1973 
 Match Study was a joint statistical project among the Social 
 Security Administration, the IRS, and the Bureau. It involved 
 matching persons in the March 1973 CPS with their survey data, 
 social security earnings and benefits, and a small amount of 
 Federal income tax data. 
 
 The purposes of the project were to use the administrative 
 data to determine the correctness of the survey data and to 
 ascertain any biases in the Bureau's statistics. The study was 
 designed also to use the administrative data to correct misreport- 
 ing of the survey information and to augment the survey with 
 
ASPECTS OF THE CENSUS BUREAU 
 
 data the respondents were not asked to provide. This type of data 
 combination can be used as inputs to models which simulate the 
 effects of a new tax transfer program or changes in -existing 
 programs, to develop corrected gross and net income distribu- 
 tions, and to test and evaluate various statistical matching 
 procedures. 
 
 Special operating procedures were instituted at both Social 
 Security and Census to insure the confidentiality of the shared 
 information. The laws and regulations which govern the agencies 
 are explicit in the restrictions on such shared information. The 
 Census Bureau hired, as temporary employees, a small group of 
 Social Security personnel. Both they and the regular Bureau 
 employees were sworn to uphold the confidentiality of the 
 information under penalty of law. Also, a special Social Security 
 commissioner's decision had to be obtained for the project. And 
 like the IRS, the Social Security Administration placed certain 
 conditions on the use and processing of its information. 
 
 Student: Just out of curiosity, how quickly do you think the 
 hammer would fall if Federal agencies tried to affiliate or link 
 data on individuals? 
 
 Mr. Barabba: Before the agreement could be made, because 
 somebody in the Bureau who didn't like it would call a friend of 
 his who would call a Congressman. It would be instantaneous. 
 
 We have a very open system in the Government, and 
 particularly in the statistical area. All this dialogue would take 
 place in open meetings. We would sit down with every one of our 
 advisory committees and go through a very similar dialogue to 
 what we're having right here. The American Marketing Associa- 
 tion might say, "That would be the most important thing you 
 ever did." But the American Statistical Association might say, 
 "It's not statistically meaningful." They might argue about the 
 methodology we're using. And the confidentiality advisory 
 committee might tell us that we were out of our minds. 
 
 10 
 
SOCIAL OBSERVATORIES THROUGH A GLASS, LESS DARKLY 
 
 Vincent P. Barabba, Director, U.S. Bureau of the Census 
 
 "By future shock I mean the disorientation and decision overload produced 
 by high-speed change .... The accelerated arrival of the future is forcing millions 
 of us ... to live at a faster tempo than our nervous systems can tolerate. It subjects 
 us to dangerous overstimulation. It predisposes us to disease. It creates psychological 
 stress. And it sometimes leads to a breakdown of our capacity for rational 
 
 decisionmaking." 
 
 Alvin Toffler, author of Future Shock, 
 from interview in J. Walter Thompson 
 house organ, "JWT World." 
 
 Millions of men and women today are in this situation. The 
 decisions they face are coming at them faster and faster- 
 decisions which affect more people, money, land, and resources 
 than ever before. 
 
 The accelerated pace of decisionmaking has made it 
 uncomfortably clear just how interdependent we are as indi- 
 viduals and organizations in society. One decision by A affects B's 
 thinking, and since B has practically no time to consider 
 alternatives, he acts, affecting C . . . . 
 
 One of the effects of the increased tempo of decision- 
 making is the truncated time available to check the accuracy of 
 information. Often the demand for action is so great that data are 
 hastily assembled and used and are simply assumed to be 
 adequate. And, unfortunately, data are often outdated or incor- 
 rect, or both. In today's world, decisions— whether at the 
 domestic household, corporate headquarters, or Federal Govern- 
 ment level— need to be made quickly, confidently, and based on 
 accurate data. 
 
 It has been said that we live in an age of rising expectations. 
 It can be added that when the tempo of achieving these 
 expectations slows, or the path toward them becomes ill-defined, 
 society becomes restless. We expect sound decisions from leaders 
 in the public and private sectors to sustain the momentum of 
 progress. And our highly effective mass communications system 
 ensures that the results of decisions become widely and quickly 
 known. 
 
 Mass communications have achieved something else. They 
 have conditioned the individual to encounter statistical infor- 
 mation as part of his everyday experience. 
 
 Statistics often appear in commercials ("Three out of four 
 doctors. . ."). Television programs live or die according to the 
 Nielsen ratings. We are accustomed to the Lou Harris, Elmo 
 Roper, and George Gallup organizations sampling public opinion 
 on the issues of the day. The rise and fall of both the Dow-Jones 
 Industrials each day and the unemployment figures each month 
 give us indications of the health of the economy. An earthquake 
 in Latin American registers 6.8 on the Richter Scale, and we 
 know it was a sizeable geological event. And if the local weather 
 report says the probability of precipitation is 30 percent today 
 and 90 percent this evening, we gauge our activity for the day 
 accordingly. 
 
 But the familiar epigram "A little knowledge is a dangerous 
 thing" was never more true than when applied to statistics. The 
 public's casual familiarity with statistics, plus its rising expecta- 
 
 tions of well-being, have joined to create a situation in which the 
 man on the street wonders why his leaders can't improve the 
 quality of life and better identify and solve today's problems. 
 
 An often heard complaint is "We can put a man on the 
 moon, but we can't set up decent urban transportation. . . 
 straighten out the welfare situation . . . build enough sewage 
 treatment plants. . . ," etc. 
 
 At the same time, measurement of social conditions is 
 inhibited by the public's overestimation of the abilities and 
 applications of computers, leading to vague fears of an ominous 
 national data bank which would store every fact of their personal 
 lives for instant retrieval by any Government agency requesting 
 information. For this fear to be realized, the United States would 
 have to abrogate not only current laws but its entire democratic 
 tradition. 
 
 A refined version of this fear is that increased knowledge 
 about society in general will lead to the ability to manipulate the 
 public mind— a fear that discounts both society's increasing 
 sophistication and its increasing resistance to persuasion. 
 
 Need for Social Reporting 
 
 The realization that a method of social reporting is needed 
 has been growing for some time. In the mid 1960's a welter of 
 proposals began to appear. The National Commission on Tech- 
 nology, Automation, and Economic Progress; a special panel of 
 41 top social scientists established by the Secretary of Health, 
 Education, and Welfare; the Presidential Commission on Federal 
 Statistics; bills offered by legislators; and many private individuals 
 and organizations— all proposed approaches to, or the rationale 
 for, a system of social reporting. 
 
 One of the more interesting proposals came from Prof. 
 Harold Lasswell of Yale, who called for the establishment of a 
 system of social observatories to provide data and analysis for 
 sound decisionmaking. He reasoned that the computer revolution 
 had made it possible to more fully realize the potential of two 
 important underutilized areas— the growing number of skilled 
 data gatherers around the country and the enormous data base 
 already collected and tabulated, but to a large extent lying fallow. 
 
 Lasswell advocated continuous decision seminars which, 
 using computers, would examine the possible effects of proposed 
 goals and strategies on various segments of society. Such studies 
 would operate in a context including the immediate, inter- 
 mediate, and remote future as well as the recent and remote past. 
 
 11 
 
SOCIAL OBSERVATORIES ... THROUGH A GLASS, LESS DARKLY 
 
 But, Lasswell did not look on such seminars as a panacea. 
 He said it was not necessary for them to present an exhaustive 
 picture, but if the facts involved could be known with a greater 
 degree of precision and certainty, the seminars would accomplish 
 their purpose. 
 
 Lasswell's desire, basically, was to put social observation on 
 a scientific footing. Certainly, the work of social scientists 
 committed to a systematic approach to measurement should be 
 accepted with the same degree of credibility as that of physical 
 scientists and not automatically relegated to a lesser standing. 
 
 A good example of this kind of thinking appeared in a New 
 York Times editorial early in 1974, which decried the lack of 
 analysis of the nature and extent of suburban expansion and its 
 results. In part it read: "Except for some sociological studies, 
 there is little except empirical observation of the chief American 
 environmental phenomenon of the century." Why is the vast 
 amount of work done by social scientists on the subject of 
 suburban growth tossed aside so readily? Is it because social 
 scientists have not agreed upon a standardized scientific ap- 
 proach? 
 
 Indeed, social scientists do not agree. Many would like to 
 operate as their colleagues do in the physical sciences. But there is 
 also a body of opinion among social scientists that their job is to 
 raise questions, not to answer them. 
 
 The social science-physical science comparison is tempting. 
 It conjures up a picture of social scientists manning an observa- 
 tory, analyzing data, and solving problems just like astronomers 
 at Mount Palomar or Kitt's Peak observatory. 
 
 But that is a misleading picture. It is one thing to accurately 
 measure the distance from Mercury to the Sun and quite another 
 to determine that a complex mix of social factors is at the point 
 in a given neighborhood that more cases of venereal disease may 
 be expected in the future. 
 
 Perhaps a much closer analogy would be that of astrono- 
 mers examining various theories of how our solar system 
 originated. While none of the theories will ever be proven beyond 
 a doubt to be correct, their formulation and the discussion they 
 spark help to solve many other unknowns and to define areas of 
 future observation. 
 
 In this case, the social scientist has an advantage over his 
 counterpart in astronomy. Like the astronomer, he can examine 
 the present and the past and formulate a theory. But, unlike the 
 astronomer, he can see it put to practical use by making data 
 available to decisionmakers to help avoid, modify, or duplicate a 
 given situation. And then, by monitoring and evaluating data to 
 measure change, he can give an indication to the decisionmakers 
 of the success or failure of their course of action. It is this ability 
 which gives purpose to the whole concept of the social 
 observatory. 
 
 What Is a Social Observatory? 
 
 At this point, we enter the semantic fog surrounding social 
 reporting. What is a social observatory? One definition is that it is 
 an organization which provides detailed demographic and social 
 intelligence, not of individuals but of statistical aggregates. These 
 aggregates in turn allow for the efficient conduct of national 
 social and economic programs and the coordinated planning of 
 effective domestic policy. 
 
 Another definition— consistent with the first— is that a 
 social observatory is an instrument which provides the means for 
 developing measures of the seeming elusive quality of life. These 
 
 12 
 
 measures are generally referred to as social indicators. 
 
 What is a social indicator? The definition of the term is 
 difficult and is only beginning to take form. The Center for 
 Coordination of Research on Social Indicates has established this 
 working definition of social indicators: "We take them to be 
 statistical time series that measure changes in significant aspects 
 of a society." 
 
 Another way to place social indicators in context would be 
 to look on them as fulfilling the same role as a blueprint. A 
 blueprint does not build a structure but provides the best 
 available information to those at the construction site whose job 
 is to put up the building. No blueprint can replace the human 
 evaluation and judgment at the site, factors which make the 
 building possible. In the same way, social indicators will provide 
 the best available information to decisionmakers. 
 
 What social conditions should be measured and how to 
 measure them is also a difficult question. One school would have 
 social reporting take the form of a system of social accounts, 
 much as our system of economic accounts reports on the health 
 of the Nation's economy. The rationale is that just as there is a 
 GNP— Gross National Product— there could be a measure of 
 GNWB, or Gross National Weil-Being. Like the analogy of the 
 physical science observatory, this is tempting because it sounds as 
 if it would result in a measure which is easy to grasp. 
 
 But this is too convenient a comparison. Our economic 
 accounts report on a structured, quantifiable entity— the U.S. 
 economy. Granted, that economy is an awesomely complex 
 thing, but the myriad elements of each part of the economy- 
 retail, wholesale, service, manufacturing, mining, transportation, 
 and construction— combine in standard building blocks to make 
 up the total American marketplace 
 
 The advantage frequently cited in relation to the GNP is 
 that all the elements can be reduced to a common denominator— 
 the dollar. This also is cited as a drawback in that the GNP treats 
 a dollar's worth of bubble gum the same as a dollar's worth of 
 penicillin. There is no effort to measure social utility. Critics say 
 there is inadequate recognition of the negative aspects which are 
 created along with the positive— for example, the price paid in 
 pollution for the manufacture of a given item. 
 
 It also should be pointed out that the system of national 
 accounts took years of effort by many brilliant economists to 
 construct and refine into reliable, consistent measures. Yet, that 
 system is not without its critics and advocates of change, in spite 
 of its wide recognition as the premier indicator of the Nation's 
 economic health. 
 
 While much of the data in the social field is also 
 quantifiable, no successful system has yet been proposed to 
 combine these factors into meaningful totals. How do you 
 measure the well-being of some 210 million individuals with 
 differing aspirations and talents? How do you develop an index 
 which will reflect happiness, respect, or dignity? 
 
 Work is underway to develop a series of social indicators 
 aimed at measuring these facets of the quality of life. These 
 efforts assume that no single GNP type of accounting— even if it 
 were developed— would be widely useful. 
 
 Defining Social Concerns 
 
 The Office of Management and Budget in the Executive 
 Office of the President recently published Social Indicators, 
 1973, which identifies eight major areas of social concern— health, 
 
SOCIAL OBSERVATORIES ... THROUGH A GLASS, LESS DARKLY 
 
 public safety, education, employment, income, housing, leisure 
 and recreation, and population. While this publication is valuable 
 because it gathers a large amount of useful data into one place for 
 the first time, much of it from the Census Bureau, it does not 
 present true indicators of how the population relates to each of 
 the areas of social concern. The introduction makes that clear: 
 
 At the present time, not enough is known about the cause 
 and effect of social conditions to develop such ideal 
 indicators. Rather, the indicators presented in this publica- 
 tion represent simply a first step toward the development 
 of a more extensive social indicator system. 
 
 Developmental work on social indicators is not confined to 
 the United States. A special committee of the Organization for 
 Economic Cooperation and Development (OECD) recently com- 
 pleted 3 years of effort to codify a list of social concerns. The 
 basic criteria were that the concerns must be common to most of 
 the 23 member nations and that quantifiable measures could 
 reasonably be expected to be developed. 
 
 The OECD list also identifies eight major areas— health, 
 individual development through learning, employment and qual- 
 ity of working life, time and leisure, command over goods and 
 services, physical environment, personal safety and the ad- 
 ministration of justice, and social opportunity and participation. 
 But these eight areas are not the ones for which social indicators 
 are expected to be developed. They are simply general headings 
 for the 24 fundamental social concerns which make up the list, 
 14 of which have subconcerns. 
 
 For example, under the broad area of Employment and 
 Quality of Working Life are 3 of the 24 fundamental social 
 concerns: (1) The availability of gainful employment for those 
 who desire it; (2) the quality of working life; and (3) individual 
 satisfaction with the experiences of working life. This third item 
 is divided into several subconcerns that include working condi- 
 tions; career prospects; and supervision, autonomy, and job 
 challenge. 
 
 This list of 24 social concerns provides the framework for 
 the next phase in the OECD program which is to develop a 
 system of social indicators or, as the introduction to the list puts 
 it, an agenda for "proceeding from the general toward the 
 specific." 
 
 By the fall of 1972 it was evident that serious work was 
 proceeding on so many fronts that it was difficult for those 
 interested to keep track of the major developments in the field. 
 This situation prompted the Social Science Research Council, 
 under a grant from the National Science Foundation, to establish 
 the Center for Coordination of Research on Social Indicators 
 based in Washington. The Center is serving as a clearinghouse for 
 significant new research in the drive to establish a system of social 
 indicators. 
 
 It is important to define the relationship between social 
 observatories and social indicators. If social indicators are the 
 information tools for decisionmakers, social observatories provide 
 the materials from which to make the tools, much as iron ore is 
 converted to high grade steels from which tools are then made. In 
 this case, the iron ore is raw data, and the high grade steels are the 
 gathering, tabulation, and first- level analysis of data. 
 
 In other words, without social observatories there can be no 
 social indicators. 
 
 Some social scientists point to the gaps in our society's 
 knowledge about itself and say there is an even bigger job to be 
 
 done than most of us realize. But the United States is better off 
 in the field of social statistics than the rest of the world. Its 
 recordkeeping was specified in the instrument which established 
 the organization of the Nation, the Constitution. 
 
 Individual Measurements 
 
 In a way, it is easy to overlook the value of the decennial 
 census simply because we take for granted its unbroken record of 
 190 years of data gathering. To us it is not unique. But the 
 French statistician Moreau de Jonnes noted in the last century 
 that the United States presented a phenomenon without parallel 
 in history: 
 
 ... a people who instituted the statistics of their country 
 on the very day when they founded their government, and 
 who regulated by the same instrument the census of 
 inhabitants, their civil and political rights, and the destinies 
 of the Nation. 
 
 Much newer, but an invaluable complement to the decen- 
 nial census, is the Current Population Survey (CPS). Taken every 
 month since 1940, through wars, recessions, booms, and periods 
 of great social change, this survey provides fast, up-to-date 
 measures of many social trends^family relationship, education, 
 mobility, and income, to name a few. Its value to users has been 
 great because the CPS data, month in and month out, have been 
 consistent, using as it does a comparable statistical universe and 
 quality standards. 
 
 Together, these two— the decennial census of population and 
 housing and the Current Population Survey— form the under- 
 girding for social measurement in the United States. 
 
 Of value too is the other, less well-known work of the 
 Census Bureau, including a number of censuses taken at 5-year 
 intervals: the census of agriculture, the census of governments, 
 and the economic censuses, which cover manufacturing, retail and 
 wholesale business, service industries, construction, mineral in- 
 dustries, and transportation. The government and economic areas 
 are supplemented by a comprehensive series of surveys, some 
 taken at weekly intervals, to fill in the 5-year gap between 
 censuses, much as the CPS does for population characteristics. 
 
 The Census Bureau could serve as a social observatory in 
 the broadest sense, having been a supplier of unbiased data for 
 almost 200 years. Since 1790 it has been the cutting edge in the 
 development of social measurement, a role it is continuing to 
 perform. 
 
 The decennial census has grown from little more than a 
 head count to an enormous reservoir of general statistical 
 information, an evolution that occurred to keep up with the 
 increasing demand for data. 
 
 Manual tabulation was adequate for the first century of 
 census taking, but despite the size of the population and the 
 length of the 1880 questionnaire, hand tabulation was still going 
 on when time came to take the 1890 census. A former census 
 employee, Herman Hollerith, invented electrical-mechanical tabu- 
 lation of information through holes punched in paper cards, a 
 system 100 times faster than manual methods. By doing so, 
 Hollerith added a new dimension to census data: timeliness. 
 
 One of the Bureau's most valuable assets is the contract of 
 trust it has built up through the years with the American people 
 regarding the confidentiality of answers recorded on census 
 questionnaires. What had been tradition and internal regulation 
 
 13 
 
SOCIAL OBSERVATORIES ... THROUGH A GLASS, LESS DARKLY 
 
 became public law at the turn of the century, guaranteeing this 
 confidentiality. This extends even to denying requests, for data 
 about individuals from other Government agencies. Significantly, 
 it has never been necessary to prosecute a Bureau employee for 
 revealing confidential information. 
 
 The gap between decennial censuses has long been recog- 
 nized as a problem. In 1940, the Bureau pioneered sampling 
 theory, in which the answers of a scientifically selected sample of 
 the population can be extrapolated to come very close to the 
 answers that a full census would provide, and at a fraction of the 
 time and cost. This technique has allowed a monthly survey of 
 population trends to supplement the data base of the decennial 
 census. 
 
 By 1950, another data-handling problem was shaping up. 
 The size of the U.S. population was such that even punchcard 
 processing meant too long a time span between taking and 
 publishing the census. The Bureau supported research and 
 development of the first electronic computer put into use for 
 mass data handling, UNIVAC 1. 
 
 More recent developments have allowed census tabulation 
 to bypass the punchcard stage, putting information from ques- 
 tionnaires directly into computers through the medium of 
 machine-readable microfilm. This method allows easier access and 
 less expensive storage since the bulky original paper forms can be 
 destroyed once the census is tabulated. 
 
 The 1970 census was largely a mailback effort, meaning 
 that most households filled out their own form and mailed it 
 back without seeing the traditional enumerator. This represented 
 large savings in time and manpower. For the mailback method to 
 work, however, a massive effort by Bureau geographers was 
 needed to produce standardized maps of all urban areas in the 
 Nation. This was done so that each address could be correctly 
 allocated to the proper geographic location. This work led 
 directly to the latest breakthrough in data handling. 
 
 Potential Use 
 
 For some time it has been realized that in one major field 
 the great potential use of census data was largely untapped, just 
 as Prof. Lasswell noted— the local level. Most problems exist and 
 are acted upon at the local level. Decisionmakers at this level did 
 the best they could with what they had, but their uses of census 
 data were severely limited. This was either through lack of 
 knowledge about what census data could do to help or because of 
 lack of time and staff to utilize census data in traditional forms. 
 Added to these factors was the problem that locally gathered data 
 were not compatible with census data. The tools existed, but 
 access to them was practically nonexistent. 
 
 In the fall of 1966, the Bureau started its first small-area 
 research and development project, the New Haven Census Use 
 Study. Basically, this was an attempt to get the more than 60 
 agencies in the New Haven metropolitan area to cooperate in 
 identifying their common statistical needs, pooling their data, and 
 developing a computer system to aid in problem solving. 
 
 The results of this project and others have been exciting 
 and have had significant impact. Tools have been developed 
 which are allowing data gathered locally to be linked with census 
 data. This combination has endless applications including more 
 effective government, environmental efforts, transportation, law 
 enforcement, education, health care, location of new facilities, 
 and market analysis. Continuing followup projects in Los 
 
 Angeles, Indianapolis, and San Francisco are refining these tools. 
 
 The tools developed include GBF/DIME— the Geographic 
 Base File, with Dual Independent Map Encoding, and 
 ADMATCH. GBF/DIME system creates a detailed computerized 
 source map which serves as a base for quickly and accurately 
 locating addresses, calculating area, measuring distance, etc. 
 ADMATCH is a group of computer programs to match addresses 
 to respective area units— street segments, blocks, tracts, etc. This 
 permits two or more data sources to be aggregated by the same 
 geographic unit. In this way, different data sources are linked, 
 allowing examination of the relationship between two or more 
 variables simultaneously. In other words, it is now possible to 
 compare data for school districts with those for police precincts 
 or with those for fire districts, even though all the local areas have 
 different boundaries. 
 
 These sophisticated computer tools, while greatly aiding 
 local problem solving, do not breach standards of confidentiality. 
 Like more traditional data formats, they are statistical ag- 
 gregates—information about groups of people, not individuals. 
 
 The full potential of locally gathered data may be realized 
 in one statement of fact: there are nearly 80,000 local govern- 
 ment jurisdictions— an overlapping complex of political bound- 
 aries, fire, sewer, school, soil conservation, water, air pollution, 
 law enforcement, and other districts. Not counting school 
 districts, there are more than 17,000 local jurisdictions in the 
 Nation's metropolitan areas alone. One of the most unsettling 
 projections in the Report of the Commission on Population 
 Growth and the American Future (Vol. 4), 1972, is a continued 
 proliferation of local jurisdictions. 
 
 The impact on decisionmaking of the ability to blend local 
 data and census data into common area units and to transform 
 these data into information, using quickly made and inexpensive 
 maps and charts, cannot be overestimated. 
 
 Punchcards, methods of maintaining confidentiality, sam- 
 pling theory, the computer, the mailback census, the geographic 
 base file— all have had a profound effect on data gathering and 
 handling. 
 
 However, a social observatory's value is greatly diminished 
 if all it does is gather data. It must disseminate information on a 
 timely basis if it is to be a vital, current social instrument, rather 
 than only of historic import. This has been realized at the Census 
 Bureau, where increasing efforts on behalf of data users have been 
 made during the past decade. This policy has been summed up in 
 these words: "The providers of data give us the wherewithal, but 
 it is the users who give us purpose." 
 
 This thrust is not simply aimed at creating more data but 
 at transferring data into information which can be put to use by 
 narrowing the range of uncertainty in the decisionmaking process. 
 Some of the user-oriented steps taken at the Bureau include: 
 
 A. Greatly expanding the publication program of the 1970 
 census. More than twice the number of pages of data 
 were issued in 1970 compared to 1960, and those pages 
 each contained about 20-percent more material. 
 
 B. Publishing data in new formats. In addition to the 
 traditional printed publications, statistical tables were 
 issued on microfilm and michofiche. 
 
 C. Making computer tapes available to the public. For the 
 
 first time, tapes covering units from city blocks to the 
 State level were sold at cost. These usually were 
 available before the printed reports were published. The 
 
 14 
 
SOCIAL OBSERVATORIES ... THROUGH A GLASS, LESS DARKLY 
 
 availability ot these tapes sparked the establishment of 
 more than 175 Summary Tape Processing Centers 
 around the Nation. These centers have been set up by 
 public and private organizations so that data users with 
 common interests can share the cost of operations on a 
 cooperative basis. Special tabulations have been put 
 within the reach of all who need them, quickly and 
 inexpensively. Once again, confidentiality has been 
 maintained, with individually identifiable data being 
 suppressed as the tapes are prepared. 
 
 D. Establishing permanent user-oriented facilities within 
 the Census Bureau. This effort has two main parts— the 
 Census Use Study, which develops innovative uses of 
 census data, and the Data User Services Division. The 
 division's charge is to improve the usefulness and 
 delivery of census data to meet the changing needs of 
 the data user community, with special emphasis on 
 service to local areas. 
 
 E. Setting up an intensive conference, seminar, and work- 
 shop program. This program is aimed at generating new 
 data users and helping those who are already familiar 
 with census data. Thousands of persons in every State 
 have participated to learn the spectrum of census data 
 products and their application. 
 
 The development of this user-oriented outlook stems from 
 the recommendation of the Bureau's eight advisory committees. 
 These committees form the key feedback vehicle from the user 
 community. Staffed by top men and women in their fields, the 
 committees represent a broad spectrum of public and private data 
 users. These Census Advisory Committees are: 
 —Committee on Agricultural Statistics 
 —Committee of the American Economic Association 
 —Committee of the American Marketing Association 
 —Committee of the American Statistical Association 
 —Committee on Population Statistics 
 —Committee on Privacy and Confidentiality 
 —Committee on Small Areas 
 —Committee on State and Local Government Statistics 
 
 These committees assist the Bureau to meet the challenge of not 
 only doing things right but ensuring the right things are done. 
 
 In its role as a social observatory, the Census Bureau will 
 continue to reconcile what is available with what is needed. It will 
 continue to provide the data base for social reporting by 
 expandingand refining its data products. This is a familiar role for 
 the Bureau-it does not make policy but provides the right 
 statistics for policymakers to help them act intelligently. 
 
 Future development appears to lie toward a dual concept of 
 social reporting-national and regional on the one hand, and local 
 
 1 Since this statement the Bureau has established two additional 
 committees: Census Advisory Committee for the Black Population for the 
 1980 Census and Census Advisory Committee for the Spanish-Origin 
 Population for the 1980 Census. 
 
 on the other, based on compatible methodologies. Schematically, 
 the data flow is represented in the chart. 
 
 The interchange between the national and local data flows 
 would be made possible by use of a common set of basic 
 geographic elements, tailored to the needs of the areas concerned, 
 yet made of compatible building blocks. 
 
 Once again, it is obvious that without social observatories, 
 there would be no social indicators. 
 
 The Census Bureau is ready to aid all levels of decision- 
 making and is unique in that its products come in building-block 
 sequence— street, block, enumeration district, tract, town, city, 
 county, State, regional, and national. 
 
 Social reporting is an idea whose time has come. Its shape is 
 unknown at this juncture, and it will be some time before a 
 meaningful system of social reporting is in general use. But the 
 first developmental steps have been taken. 
 
 Problems 
 
 At this stage, there are three main problems. One is 
 definition. What do we want to measure? Given the incredibly 
 complex interaction of the elements of society, there is an 
 infinite variety of activities, accomplishments, and needs to be 
 measured. It must be realized that what we don't know will 
 always be greater than what we would like to know. 
 
 This leads to the second problem, setting priorities. Given 
 an infinite number of items to measure and finite resources with 
 which to measure them, the hard business of choice must be 
 faced squarely and a beginning made. Initial choices will almost 
 certainly not be comprehensive. Indeed, the list will constantly 
 need revision. Since the cornerstone of social reporting is 
 measurement over time, changes must be made only when clear 
 need is established and agreed upon. 
 
 The third problem is utilization, something to which the 
 Census Bureau can contribute, while definition and priorities are 
 developed. It is doing so by making its data products more widely 
 known, more usable, more accessible, and available in a reduced 
 amount of time. 
 
 There are many interested groups working toward a system 
 of social reporting. The Census Bureau is ready to participate 
 with all of them by making its past, current, and future 
 information available, as well as its capability and understanding 
 of data collection, developed as it accomplishes its mission as fact 
 finder for the Nation. 
 
 This is an exciting time to work in the social observatory 
 field. The stakes are high. As Alvin Toffler said in his interview 
 with"JWT World." 
 
 Super-change, and the decision crisis it is provoking in our 
 culture, will compel us to think about change in totally new 
 ways. Those who cannot will fall by the wayside— casualties 
 of the accelerative thrust, victims of future shock. 
 
 The role of the social observatory is nothing less than to 
 minimize the number of those victims. 
 
 15 
 
SOCIAL OBSERVATORIES ... THROUGH A GLASS, LESS DARKLY 
 
 POPULAT/Q/V 
 
 OBSERVATORIES 
 
 ST LEVEL ANALYSIS OF 
 iING AND EVALUATION OF 
 
 PROGRAM 
 
 OPULAT 
 
 s\s^+ 
 
 PRIMARY DATA FLOW 
 
 MONITORING & EVALUATION 
 
 EFFECTS OF DECISIONS 
 ON POPULATION 
 
 16 
 
A DISCUSSION OF PRIVACY 
 
 Vincent P. Barabba, Director, U.S. Bureau of the Census 
 
 "The point I hope I made in my paper was that information maintained to describe 
 or characterize specific individuals requires a whole different set of rules and regulations 
 than aggregate information released only in total and not related to any specific indi- 
 vidual." 
 
 Vincent Barabba 
 
 Mr. Barabba: The debate over the individual's right to privacy can 
 really get out of hand. Recently, I testified before Congressman 
 William S. Moorhead's Subcommittee on Foreign Operations and 
 Government Information concerning the Goldwater-Koch Bill 
 and others which deal with the responsibilities of those who 
 collect or maintain individually identifiable information. 
 
 The Goldwater-Koch Bill is quite good as long as people 
 realize that there are significant differences between agencies that 
 collect data from individuals for statistical purposes and those 
 which maintain dossiers for administrative purposes. 
 
 The Goldwater-Koch proposal is designed to regulate the 
 collection and use of data by Federal agencies. It incorporates 
 most of the central guidelines proposed in the 1973 HEW volume 
 Records Computers and the Rights of Citizens. The book was a 
 landmark example of a Government agency's recommendations 
 for regulation, rather than those of a private agency. 
 
 Let me give you an example of some of the quasi-political- 
 philosophical confusion that arises over the privacy issue. 
 Moorhead's Subcommittee was talking about a provision of the 
 bill which states that the Social Security number may only be 
 used by the Social Security Administration. Congressman Paul 
 McCloskey, a member of the Committee, indicated that some 
 district attorneys reported the use of the Social Security number 
 was very helpful in tracking down individuals who had fled a 
 particular area. Congressmen McCloskey asked a Congressman 
 who was testifying if the bill under consideration wouldn't inhibit 
 the ability of law enforcement agencies to locate people. The 
 witness responded that in the case of persons who were 
 delinquent on child support payments, and faults of that nature, 
 the authorities ought to be able to use the Social Security number 
 to get them; but he added if a person was delinquent with a car 
 payment, pursuit via the debtor's Social Security number would 
 be hard to justify, and the Social Security number should not be 
 used. 
 
 At that point I leaned over to my legal counsel and asked if 
 I had missed something! You can see why privacy is a very 
 emotional topic. It was all right to use the Social Security number 
 for things that the Congressman favored philosophically, but it 
 was quite different with regard to those he opposed. If at all 
 possible, we have to keep the privacy issue from becoming a 
 political football. The rules we lay down must be well thought 
 out and considered, with distinctions clearly drawn. 
 
 Getting a rational conversation going is a very difficult task 
 in political Washington. It's a lot easier to sound as though one is 
 protecting his constituents by saying "No" to everything related 
 to their right of privacy. It takes a person with convictions to 
 argue for the Government's need to know. Because of this 
 dilemma the Bureau is going to spend more time, as an 
 institution, trying to tell the other side of the story. If we don't, 
 we fear we may have to tell the Federal agencies and Congress we 
 
 can't get information for them anymore because we can't afford, 
 or we're not allowed, to collect it under the privacy restrictions. 
 At the very same time we testify before one committee 
 concerned about privacy and the possibility of collecting too 
 much data, we may have to go to another committee and explain 
 why we aren't able to give them all the information they feel they 
 need. 
 
 Take the education bill passed recently. There's a provision 
 that within a year of its enactment the Secretary of Commerce in 
 conjunction with the Secretary of Health, Education, and Welfare 
 shall expand the Current Population Survey to provide estimates 
 of the population by age range within poverty households. That's 
 a very, very difficult task in the first place. But, if the rule was 
 laid down that we couldn't call on people in their homes, it 
 would become impossible. 
 
 We have to continue reminding each committee of the 
 Congress what the other committees have told us we can't do. It's 
 not that the Congressmen are inconsistent. But, the committee 
 structure is such that sometimes one committee will operate 
 without others being aware of its activities. All the privacy 
 legislation being debated has little to do with census legislation, 
 which goes through an entirely different set of committees, 
 although privacy rulings will directly affect the Census Bureau. 
 Because of this system the Bureau has to keep its eyes and ears 
 open to all types of legislation being considered. Recently, in 
 reaction to the Watergate situation, Senator Lowell P. Weicker, 
 Jr., attached an amendment to the White House Authorization 
 Bill. The amendment says that if anybody in the White House or 
 a Federal agency wants to examine individual records from the 
 Internal Revenue Service, he must have an authorization per- 
 sonally signed by the President of the United States. While 
 Senator Weicker saw this amendment as necessary to protect the 
 Internal Revenue Service from future onslaughts by White House 
 personnel— and I agree in principle— because of the wording, it 
 unfortunately would also hamper the entire administrative 
 records program at the U.S. Census Bureau. 
 
 In response to amendments such as Senator Weicker's, the 
 Bureau must go to the appropriate conference committee and 
 explain its side of the story. In the case of the Weicker 
 amendment the Bureau would face a major problem, as passage 
 would deprive the Bureau of the means to produce intercensal 
 population estimates. The conference committee has to decide 
 whether to approve the amendment or kill it in committee. 
 
 The privacy debate has come about, I think, because right 
 now in the United States we are seeing revealed the abuses- 
 potential and real— of personal records. Unfortunately, people are 
 tending to confuse the meaning of expressions like "the right of 
 privacy." All of you filled out a registration form for this class. 
 You could have claimed that was an invasion of your privacy. 
 You were willing to provide information on that form, including 
 
 17 
 
A DISCUSSION OF PRIVACY 
 
 your Social Security number, in order to be able to get the credits 
 for this course. Everyday, it seems to me, aspects of this "right of 
 privacy" are compromised. 
 
 Student: Why is it even necessary to record an individual's name 
 on a census schedule 7 
 
 Mr. Barabba: There are two reasons. One is the service that we 
 perform for people who need to verify where they were at a 
 certain point in time. Only that person or a direct bloodline 
 relative can gain access to that information. 
 
 Until recently— and in some parts of the country even 
 today-administrative records such as birth certificates and 
 property tax rolls were not carefully maintained and census data 
 was the only accurate record available. Prior to 1900 Bureau 
 records were all publicly available. There has been a tremendous 
 demand for these 1900 to 1930 reports because people moving 
 into the Social Security benefits age range didn't have birth 
 certificates to prove their age. An individual's documented census 
 record will still today suffice as one of the proofs of age and 
 nationality for Social Security and passports. 
 
 The other reason is that when we do some of this matching 
 activity discussed earlier, we want to call back on individuals. In 
 other words, if a program needs additional information on 
 disabled persons, we have to be able to go into our files and find 
 the addresses of people who are identified as being disabled. 
 These persons may have moved, so we need to know the names in 
 order to discover their location. We would attempt to use these 
 people as part of our sample, though it would be voluntary 
 whether they responded. 
 
 The important thing is that for the 1970 census we 
 separated the name of the individual from his or her responses. 
 The name never goes into the computer. Individual names can be 
 found only on a microfilm copy of the original questionnaire, and 
 the copy is kept in a secure file for the specific purposes I 
 mentioned. I am aware of no other records. 
 
 Student: I have the feeling that privacy is an emotive issue. 
 Perhaps for comparative and persuasive purposes you should 
 offset it with another emotive term such as "secrecy." 
 
 Mr. Barabba: That's interesting, because the Swedes have what 
 they call the "publicity principle." In Sweden, until very 
 recently, nothing was secret; absolutely nothing. Any public 
 record was open to the general public for review. Tax records, 
 everything, were there for anyone who wanted to see them. 
 
 The argument in favor of this system is that one need not 
 worry about the Government ever doing something behind closed 
 doors. This principle evolved over several centuries, and goes back 
 to the 1600's when Sweden's church kept a list of all the people. 
 Then, for a while, the Government retained the church to keep 
 the list and now the Government does this itself. As Sweden is a 
 socialized society, the individual's benefits are derived from a 
 system where the Government needs to communicate directly 
 with its constituents and needs considerable detailed information 
 for planning. So, when you move, you work very hard to keep 
 the Government apprised of where you are because your whole 
 benefit package is tied to whether you're on the list. But that list 
 and everything on it is generally publicly available. There have 
 been attempts to restrict some of the information. It's a lot easier 
 to maintain such a policy in a society which has grown up with 
 
 that tradition than it would be in the United States with its 
 traditions of individuality. 
 
 Let me change the topic and ask for your reactions to the 
 general theme of my "Right to Privacy" presentation. 
 
 Student: I felt that you had some kind of an ax to grind while at 
 the same time you were trying to present objective information. I 
 felt you were justifying the existence of the Census Bureau and 
 what it does and what it needs to do, which may be misin- 
 terpreted as an invasion of privacy. 
 
 Student: I like the word "justify" because the allocation of 
 resources and continued socioeconomic growth and development 
 demand you know what trends are occurring to ensure effective, 
 efficient decisionmaking. 
 
 Student: I have the feeling that you didn't just sit down and say, 
 "I'm going to write what the Census Bureau does." It reads more 
 like a reaction to forces that have been pressuring you. 
 
 Mr. Barabba: I think that's true. I probably spend 20 percent of 
 my time grappling with the privacy issue, and to that extent it is a 
 reaction. However, we're more than satisfied with our operating 
 law, which is quite clear about what we can and can't do, and it's 
 the toughest existing privacy legislation on the books today. It's 
 so restrictive that other agencies of Government have a hard time 
 living with us because we cannot let them look at the individual 
 responses. We'll only give out statistical aggregates, unless we 
 collect the information under another agency's authority and tell 
 those being interviewed just that. 
 
 Student: Then, is the problem that some people, either the 
 Congress or the general public, are uninformed as to how clearly 
 delineated both in the statutes and in practice your treatment of 
 the respondent's privacy actually is? 
 
 Mr. Barabba: That is correct. We have the problem of the 
 Congress creating new legislation which, in a sense, is to be 
 superimposed over the existing census confidentiality statute. In 
 the cold of the energy crisis we argued against proposed 
 legislation to allow the Federal Energy Office use of individually 
 identifiable Bureau records of corporations and business enter- 
 prises. This caused us considerable concern, as our governing 
 statute states quite clearly, and so we tell those we interview, that 
 business records are only to be seen by Bureau employees. In 
 fact, there was a Supreme Court case which led to the Congress 
 enacting specific legislation to protect business reports. If other 
 agencies were able to use the individually identifiable records in a 
 way that could be harmful— or interpreted as harmful— to 
 businesses, our statistical accuracy would go out the window 
 overnight. 
 
 Student: As a probation officer I'm aware of the many calls our 
 office gets from outsiders who are seeking information about 
 individuals. I'm concerned that any information we give out may 
 be harmful to the individual in question. I feel the same way 
 about your census data. That is, can the information you collect 
 eventually reach another agency in some sort of individually 
 identifiable form? 
 
 Mr. Barabba: The point I hope I made in my paper was that 
 information maintained to describe or characterize specific 
 
 18 
 
A DISCUSSION OF PRIVACY 
 
 individuals requires a whole different set of rules and regulations 
 than aggregate information released only in total and not related 
 to any specific individual. I think the rules now being developed 
 as restrictions on personal data banks must be tight. I also think 
 people at the local level must maintain the same attitude about 
 their data as we at the Federal level. 
 
 We are concerned, as well, with the responsibility of local 
 officials as regards the privacy of individuals. An ideal privacy law 
 would reach down to the local level in its protection of 
 individuals, because the closer you are to people the more chance 
 there is that you can hurt them directly and individually. If you 
 want to know about an individual in a town of 10,000 people, 
 you are not going to run back to Washington and go through his 
 census questionnaire. You're going to want to go to the local 
 police department or the neighborhood credit bureau. You'll just 
 go to these local files and pick out the information you want. 
 
 Student: I tend to think that most people are honest and don't 
 have much to fear as far as information collected by the Census 
 Bureau goes. On the other hand, I think people are and should be 
 concerned about the individually identifiable information col- 
 lected and exchanged by corporations, banks, and credit agencies. 
 
 Mr. Barabba: That is another interesting problem. One of the 
 pieces of legislation being proposed says that, if a Federal unit 
 holds information about an individual, that person must be 
 informed. If you plan to release information about an individual 
 for other purposes than those for which it was requested he 
 would be told that as well. Some would even say that you have to 
 tell a person every time you change his record. Additionally, the 
 company, bank, or whatever, would have to maintain a system 
 that allows retrieval of all the information collected about any 
 individual. And, if it can be shown in point of fact that it is 
 incorrect, they would have to make the appropriate changes. 
 
 One thing that has to be kept in mind when we consider 
 placing these sorts of requirements on companies, or anyone, is 
 that it is going to cost more money to process individual records. 
 This increased cost is because a system that is created and 
 maintained to allow rapid access to these records is more 
 expensive to design and operate. We have to ask ourselves 
 whether it is worth those extra dollars to make certain that a 
 person's confidential relationship with a bank, agency, or 
 company hasn't been breached. 
 
 Student: In "Privacy and the Need to Know" you write about 
 mailing lists and how you view them as a slight annoyance at best. 
 Yet, I'm really skeptical about direct mail advertising, and I 
 object to the intrusion on my privacy. My belief is that, if you go 
 out and shop around for the same item, you can beat the mail 
 price. I don't touch door-to-door salesmen or junk mail. Would 
 you elaborate on your philosophy. 
 
 Mr. Barabba: Let's explore the question of whether or not the 
 sale of your names to someone who advertises by direct mail is 
 bad or good. My argument would be to determine whether your 
 individual privacy, in this case the minimum of your name and 
 address, is of greater value to society than the value to society of 
 the actions performed by the release of the information; namely 
 the distribution of goods. 
 
 Let's talk about it from a marketing point of view. What is 
 marketing? Marketing is the efficient distribution of goods, 
 
 economic goods, to a society. You have products and you have 
 people. We must now find how to get them together. 
 
 We'll just take a simple model and create three categories: 
 retail, wholesale, and sales (company sales). One of the things 
 that we know immediately is that consumers are different people. 
 They don't think the same, they don't act the same, and it's 
 seldom that you have a product that everybody sees the same 
 way. Therefore, one of the things you do in marketing is define 
 an audience; that is, find those most likely to buy your product. 
 But you've got to communicate with those people. Now, if you 
 have some products that have a certain utility, the most important 
 thing you've got to do is get it to the people who require that 
 service. After you reach those most interested or in need of the 
 product, the product itself becomes the most important aspect of 
 selling. 
 
 Then, there are some products that are of such a nature 
 that they don't fit easily into this normal distribution pattern. 
 Here you want people to know that a certain product is available, 
 and if it meets some personal desire they will purchase it. To a 
 certain extent advertising can be looked upon as an education. 
 Let's say that I've got a product and it is the kind of item that 
 appeals to a single person who is over 30 years old and makes 
 above $30,000. I ask myself if I should create a sales force to visit 
 all the wholesalers, who, will themselves, reach all the retailers. 
 And should I develop and deploy advertising that will reach a 
 specific type of person? I also wonder whether the retail shops 
 effectively cater to people with the customer profile? If so, I 
 would have part of my marketing function pretty well accom- 
 plished. I could distribute my product through the wholesaler and 
 the retailers, who increase the cost to the public by adding their 
 profit margin. 
 
 But, I can lessen the cost if I bypass the wholesaler. To do 
 this, I need the list of people who fall into the customer profile, 
 so I can go to them directly. Since we've already determined that 
 the likelihood of their wanting a product like mine is relatively 
 high, I might be doing them a service by letting them know that 
 my product exists, while selling to them at a lesser cost than the 
 wholesaler-retailer route. Perhaps this approach would also make 
 it possible for persons not to have to travel so far to purchase the 
 product, thereby saving energy as well. 
 
 The point I want to make here is that I can go into the 
 direct mail business with the best interests of the consuming 
 public in mind because what I'm trying to do is to provide them 
 with my product in the most convenient and inexpensive way 
 possible. However, the key to all this is that I've got to get my 
 hands on a good list of names. The alternative is to use mass 
 advertising, such as television. Granted, your exposures per dollar 
 on television are cheaper, but you're reaching a lot of people who 
 are not necessarily interested in the product. The manner in 
 which I'm communicating, through direct mail soliciting, is not 
 necessarily offensive. But the message might be offensive, which 
 is true of all advertising and communication. 
 
 Student: How about looking at it this way? By putting more of 
 this junk mail in the system, the mail carrier has got to carry 
 more, which puts a bigger load on the Post Office Department 
 and jacks up the rates to everybody that isn't interested in your 
 product. 
 
 Mr. Barabba: You use the term "junk mail." Let's look at it 
 another way. If I've conducted good research and I'm reaching 
 the right people and offering them a specific message that is of 
 
 19 
 
A DISCUSSION OF PRIVACY 
 
 interest to them, do you still call that junk mail? I would imagine 
 the people who learn of a product that satisfies a need don't 
 think it's junk. In fact, some people call it direct advertising. 
 
 Student: If it's not junk mail for that target group, it's mail that 
 has to be handled by the system, which takes up time and has to 
 be carried to a certain percentage of people that don't want it. 
 
 Mr. Barabba: Either I've made an assumption or I have hired a 
 survey research firm which has interviewed those people, shown 
 them the product, and asked what they thought of it. If some of 
 them told me, "I really like it," and some of them told me, "Who 
 the hell needs it?" then, in a sense, I am a "bad guy" by telling 
 everybody, even those who don't want it, about this product 
 through a mass communication medium like television? 
 
 Student: The real issue is the invasion of privacy. Personally, a 
 telephone solicitor or someone that comes to my door is invading 
 my privacy. When I need something, I can go out of the house and 
 get it. I don't want to be interrupted at someone else's 
 convenience. For me, the direct mail business is not so much of a 
 hassle, because I pick up the mail when I'm ready, and I have the 
 option of reading it or not reading it. 
 
 Student: I don't want people to know my address. I don't like 
 people to know my telephone number, only those whom I 
 choose. I have an unlisted number, but I know that number is 
 given out because I get various phone calls from people soliciting 
 things. 
 
 Mr. Barabba: I'll tell you how that is done. Let's say one wants to 
 conduct a telephone survey of Los Angeles County. He knows 
 that 30 percent of the numbers are unlisted. What he does is find 
 the number of people on each exchange. Let's say the "367" 
 exchange has got 500,000 people, and the "354" exchange has 
 got 400,000. So, we draw a probability proportionate-size sample 
 and say the 367's are going to have more interviews in them than 
 the 354's. Then, all you need is the "367" series. Starting at 0000 
 to 9999, you draw randomly from the four digits. Whether your 
 name is listed or not, the likelihood of your number turning up is 
 proportionate to the number of people on your exchange. 
 
 Student: How do they know your name? 
 
 Mr. Barabba: If they know your name, that's a different story. 
 Then somebody obtained your name and your telephone number 
 together. 
 
 Student: So again, we have some undisclosed source giving my 
 number out. 
 
 Mr. Barabba: That is correct. Today, legislation is being con- 
 sidered that would require persons using any individual's name 
 for a purpose other than that for which it was originally supplied 
 to let the individual know about it. Right now this legislation 
 would affect only agencies of the Federal Government, but some 
 would advocate that this rule should apply to the private sector as 
 well. If these sorts of rules were put into effect, the cost of 
 acquiring the list would be much greater. We, as a society, will 
 have to ask ourselves what privacy is really worth. I don't think 
 that the marketing profession has any idea of what it would cost 
 
 to institute the sort of privacy protection that I've been 
 describing. 
 
 Student: It seems that you are arguing that the cost of reporting 
 unauthorized uses of an individual's name and other personal data 
 will make it impossible to distribute goods economically within 
 our market-oriented society. It sounds like the individual doesn't 
 really have a choice about personal information being used. Is this 
 unavoidable? 
 
 Mr. Barabba: Not at all. If the legislation I've just been talking 
 about is passed into law, the individual will be very well protected 
 as far as unauthorized usage by the Federal Government goes. It 
 is already the case that individuals are protected as regards the 
 private sector and its credit reporting activities, at least to the 
 extent that misusing credit information about an individual is 
 illegal. Soon there will be legislation concerning the improper uses 
 of medical records. But whether there should be more restrictions 
 placed on our market economy in the name of individual privacy 
 is open to debate. We absolutely must weigh the alternatives. 
 
 It seems to me that we have to become aware of the fact 
 that in our society the distribution of goods is based firmly on a 
 system which requires a free flow and exchange of information. I 
 would favor a more efficient way of reaching individuals that 
 doesn't intrude on people's privacy. Remember, part of the 
 economic system is also the right to choose, the right of personal 
 selection. I don't think anyone here wants to move toward a 
 system that would offer less choice or variety in goods and 
 services, and yet that is exactly what restricting the flow of 
 information could do. So, I hope it is clear that the problem we 
 have been talking about, that is, direct mail advertising and the 
 efficient allocation of goods and resources, is by the very nature 
 of our system interwoven with the right of individual privacy. 
 
 Student: I think this is part of the problem of identifying social 
 indicators, because there is no way to measure the cost to society 
 of certain actions. I'm looking at it from the consumer's 
 standpoint. I, personally, would not know how to judge the cost 
 of giving up a little more of my privacy. 
 
 Mr. Barabba: The real cost to the consumer of the product will 
 rise or fall depending upon the most efficient communication 
 system. There's another problem. Using the direct mail method, 
 you have an idea of what your maximum sales might be based on 
 the number of people contacted. So, you produce in a quantity 
 directly related to this particular form of distribution. If you go 
 the television route, your key is to sell as many units as possible 
 at the lowest possible price. Society may not need as many units 
 as you hope to sell. Why squander limited resources, raw 
 materials in this case, to build an unrealistic quantity of products 
 when they could be used for more important things? 
 
 Student: Wouldn't it be better, if you wanted that kind of 
 message by mail, to sign up requesting it and tell them how many 
 kids you have, and if you're married; all the particulars. Then 
 other people don't have to put up with it. 
 
 Mr. Barabba: I would favor another way. Rather, I'd inform the 
 solicitor, "I have absolutely no need for this kind of product and 
 whoever sold you my name didn't know what he was talking 
 about." I'd prefer the free enterprise system work that way. What 
 
 20 
 
A DISCUSSION OF PRIVACY 
 
 I'm telling him is that the guy he bought the list from sold him a 
 bill of goods. 
 
 Let's talk about the concept of "future shock" from the 
 information point of view. One of the points relevant to this 
 group was the impact that all of this new information, and the 
 volume of information, has on individuals. How do you identify 
 that problem? 
 
 Student: I would say that's an important point. I don't see 
 change per se as the problem, but how we as individuals respond 
 to change. I think one central issue is the belief of most 
 individuals that they control their own lives to a certain degree, 
 or at least trust those people who do control the things the 
 individual can't. You have to make a decision about some things. 
 You have all this information and no time to undertake the 
 methodical, scientific consideration of the information that many 
 of us have been taught is the rational process. I think you just 
 give up eventually and just throw it all in. Voters elect 
 Congressmen who make foreign policy, and these men should 
 theoretically know about what's happening in Cyprus and all over 
 the world. You just can't know everything. I think that's one of 
 the major problems. 
 
 Mr. Barabba: It's the plight of the Ancient Mariner, you know, 
 "Water, water everywhere, nor any drop to drink." It's that single 
 piece of data that you need that really becomes most important. 
 
 Student: In terms of the cost, I'm just wondering if the decennial 
 census in some way might be outdated from a cost/utiiity 
 standpoint. 
 
 Mr. Barabba: There's a lot of discussion that takes place about 
 your question. For example, if we are interested only in the 
 number of people, there are more efficient ways on an intercensal 
 basis to estimate what the population of the United States is at 
 any given point in time than to actually go out and count it. In 
 addition to that, we can give it to you as local area data, as well. 
 
 The problem is that as soon as you tell people what the 
 count was, the next question they ask is, "What was the 
 composition of the count?" Let's say you could get age, race, and 
 type of abode that a person lives in. Is that enough? To explain 
 some phenomena, two kinds of information are enough. For 
 others it takes more. For example, to get local-area data about 
 education, a sample of the population is not as efficient because 
 we may interview only two or three hundred people in an area 
 like Los Angeles. I wouldn't be comfortable with the confidence 
 interval for a survey of two or three hundred people in a large 
 metropolitan area, particularly if the county of Los Angeles 
 wants to know the difference between Los Angeles and the rest 
 of the large cities. 
 
 At this point, large scale census-type data become more 
 important, especially if you have to break down a large universe 
 by geography. If you're interested in national aggregates, it's the 
 worst way to get information, it is so expensive and such a slow 
 process. On the other hand, the Current Population Survey which 
 measures unemployment, where timing is more important than 
 anything else— a survey we complete in 15 days— is done on a 
 sampling basis rather than by complete enumeration. 
 
 Student: What's happening in terms of getting more information 
 to the individual? We're creating so much information, yet we 
 
 seem to be talking about making information available to the 
 public official or special groups. But the individual has sort of 
 been swept away. Is there anything being done to increase the 
 information available to the average private citizen? 
 
 Mr. Barabba: During our lunch break, the representative from 
 KNX radio called and wanted to do an interview on what the 
 Census Bureau is doing at the University of Southern California. 
 We had a nice chat. So the press is certainly one means of 
 communicating with individuals. One of the things the Bureau 
 can do is to put our information in more usable formats that are 
 of general interest to the population. We've talked with the 
 library associations about considering libraries as "retail" outlets 
 for census information. We're figuring out ways of developing a 
 dialogue within the library community, so when people raise 
 questions that require census information, the librarians will 
 know how to go through census volumes and point out where 
 each kind of information is located. We also have a potential 
 project under discussion with Oklahoma State University. 
 Through their university extension program, we may attempt to 
 set up a system where their county agents are taught how to use 
 census materials, and not just the agricultural aspects. Then, in 
 each of those small communities when they have to make 
 decisions relative to public administration and things of this 
 nature, a person is available to provide the information to people. 
 I'm not too sure I can better answer your question, other than 
 to say we're going at it from an educational point of view. We 
 have developed a film about the American people called "We" 
 and another called "Fact Finder for the Nation." These films are 
 now made available to high schools, and we have a person 
 working on a week-long curriculum that a secondary school 
 teacher could use explaining the Census Bureau, the kind of 
 information that's available, and what it tells about the American 
 public. 
 
 Student: Aren't we trying to infer that, with the rapid change of 
 technology and lifestyles, to get accurate information or make 
 decisions we're going to require a higher degree of explanation 
 which implies that we're going to have to examine more and more 
 variables, and hopefully, accurately? 
 
 Mr. Barabba: Or you find ways of using the variables that you 
 already have more efficiently. Let's talk about some of the things 
 that you have to deal with in your everyday work. Isn't 
 somebody a probation officer? Let's talk about that. What are 
 some of the things you have to deal with in your job? What kind 
 of things do you have to describe about the whole concept of 
 probation or some given part? 
 
 Student: As an administrator, you certainly need to know the 
 rate of success in terms of probation. 
 
 Success in terms of the offender would be no further law 
 violations; eliminating future law violations; meaningful occupa- 
 tion, being a student or gainfully employed. If they're on 
 probation, it would be adherence to the conditions of the court. 
 Let's say they're paying restitution. Then success would mean 
 prompt and steady payments. 
 
 Mr. Barabba: Assume these are success. Which do you think is 
 the most important? If you were limited to one on which to base 
 your whole program, which one would you choose? 
 
 21 
 
A DISCUSSION OF PRIVACY 
 
 Student: If a kid goes in the first time for a narcotics violation 
 and he comes out on probation, he may do dope every day after 
 that and never get caught. That does not show. There's no 
 recidivism in this kid; you'll think he's cured. 
 
 Mr. Barabba: But if he was still doing dope, wouldn't he be giving 
 up gainful employment and other things. Thus, the variables are 
 relative to each other. 
 
 Student: Most likely. 
 
 Mr. Barabba: We haven't really even described what success in 
 probation is. If probation success was related just to recidivism, 
 you would need one variable to describe it, that being whether he 
 goes back in the "slammer." The next question you should raise 
 is whether that is really a sufficient goal to determine success. 
 And I think that anybody who's been in the business very long 
 would say no. One of the ways you can fully describe a 
 phenomenon is to make it quite simple, but that doesn't do the 
 job for you. As you go into the description of a more complex 
 phenomenon, you come to several problems: How many factors 
 are there? How do they relate to each other? Which ones are the 
 most important? That is, should different factors be given 
 different weights? 
 
 Let's say there's a big debate going on in your profession 
 relative to what success really means. One thing you could do is 
 to get people who have a good reputation in your area together 
 and have them talk about specific case studies of individuals 
 they've run into through the probation cycle and about the ones 
 that they feel are real success stories. Then you could talk about 
 those. With 10 people around a table talking about 20 cases you 
 can see how many things they agree are measurements of success. 
 
 Then you could say, 'These are all successes, now let's look 
 at the characteristics of these people along certain lines." You 
 could develop scales of occupation from low to high, gainful 
 
 employment from low to high, and education from low to high. 
 Then you could run different kinds of mathematical devices to 
 find out the extent to which any of these variables contributes to 
 the thing called success. You can actually develop a formula, and 
 you just run these things through it and say that these are 
 variables that ought to be looked for. You can determine that one 
 or another tends to be more important, because it contributes to 
 the description of success. In an area like probation, that 
 approach may be 10 times better than your current procedures, 
 and for the next decade that may be one tremendous goal. 
 
 Relating this to "future shock," with all this information 
 we receive, if we start thinking about our information needs, 
 relative to accurate description of phenomena, you could end up 
 asking for less information rather than more information. Because 
 if you, as a probation officer, ask for all of this information and 
 then some, and you in the health area ask for all of the 
 information you need, we may well find that the requirements 
 overlap and that's helpful, but a lot of time they don't and you'll 
 want a slightly different formula. 
 
 At the same time, you have an organization like the U.S. 
 Census Bureau coming out with the big thick books and, in a 
 sense, contributing to the problem rather than solving it. Now, if 
 we can come up with two or three pieces of information with 
 multiple uses, we can solve an awful lot of problems. 
 
 Student: You're saying that the more the phenomena proliferate, 
 the less we should know about them if we want to make a decent 
 decision. 
 
 Mr. Barabba: No. I'm saying that we should try to understand 
 more meaningfully each phenomenon by finding those 
 characteristics that best describe it. In a sense, you know more 
 about it because you are dealing more fully with the meaningful 
 measures rather than "thumbing through" all possible measures 
 of the phenomenon. 
 
 22 
 
THE RIGHT OF PRIVACY AND THE NEED TO KNOW 
 
 Vincent P. Barabba, Director, U.S. Bureau of the Census 
 
 Recently, while reviewing a Census Bureau project to design 
 an ideal census for the year 2000, it occurred to me that we are 
 now, in August 1974, closer to the year 2000 than we are to the 
 end of the Second World War. That's hard to grasp, because the 
 year 2000 sounds so far in the future. It also occurred to me that 
 in just 10 years it will be 1984. 
 
 Nineteen eighty-four. I'm sure as we get closer to that year 
 there will be another printing of George Orwell's powerful work 
 about what totalitarianism can do to the human spirit. Recently I 
 took my copy from the bookshelf and reread it. Even though it 
 had been a long time since I had picked it up, several things 
 remained clear in my memory. For instance, the slogans of the 
 ruling party: "War is Peace," "Freedom is Slavery," and 
 "Ignorance is Strength." Most of all, I remembered the thought 
 police monitoring a person's every movement on two-way 
 television screens, and the saying "Big Brother is watching you." 
 
 The book's very title has come to stand for anything which 
 infringes or conflicts with the natural desires and rights of the 
 individual. To put it another way, "1984" stands for dehumani- 
 zation, regimentation, and the lack of free choice. 
 
 The noted Soviet author Aleksandr Solzhenitsyn has 
 written: 
 
 As every man goes through life he fills in a number of 
 forms for the record, each containing a number of 
 questions .... There are thus hundreds of little threads 
 radiating from every man, millions of threads in all. If these 
 threads were suddenly to become visible, the whole sky 
 would look like a spider's web, and if they materialized as 
 rubber bands, buses, trams and even people would all lose 
 the ability to move, and the wind would be unable to carry 
 torn-up newspapers or autumn leaves along the streets of 
 the city. They are not visible, they are not material, but 
 every man is constantly aware of their existence .... Each 
 man, permanently aware of his own invisible threads, 
 naturally develops a respect for the people who manipulate 
 the threads. 
 
 And that's what I would like to talk about today— the 
 relationship of the natural desires and rights of an individual 
 living in our society, specifically the right of privacy, with the 
 legitimate need of society to collect information about the 
 individual, in order for society to understand itself. 
 
 The specter of government intrusion into the affairs of 
 individual citizens has always been a highly emotional subject 
 and, unfortunately, one often discussed in the framework of 
 emotional newspaper headlines. Events of recent years have 
 brought the issue into the forefront of public debate. But at the 
 same time the highly emotional character of events such as the 
 military surveillance of civilians, wiretapping, and political espio- 
 nage have tended to obscure the nature of a whole series of 
 privacy-related questions. Emotionalism has reduced this very 
 complex problem to a simplistic assertion that each man has an 
 inalienable right to privacy and no more invasions of that privacy 
 should be allowed. Yet, this dogmatic approach obscures and 
 hampers objective debate, as well as decisionmaking, concerning 
 
 aspects of individual privacy and confidentiality that collide with 
 society's need to know. 
 
 For example, in various speeches and congressional testi- 
 mony Senator Sam Ervin, a most distinguished constitutional 
 lawyer and advocate of an individual's right of privacy, has used 
 as examples of unlawful or questionable government intrusion the 
 compilation by military intelligence of dossiers on civilians, and 
 also the questions asked in the decennial census and other surveys 
 by the Census Bureau. 
 
 While I know that Senator Ervin understands the distinc- 
 tions between these activities, the context of his speech does not 
 reveal the differences between the collection of information for 
 administrative purposes and the collection of information for the 
 publication of anonymous statistics. Failing to distinguish be- 
 tween types of data collected and their purposes does much to 
 contribute to the headlines that cram my files, such as "Reversing 
 the Rush to 1984," "Who Knows? The Computer Knows," "Big 
 Brother Society Feared," and "Federal Data Gathering Like 
 Octopus." 
 
 Many of these articles mention the Census Bureau in a 
 catalog of "sinister or potentially sinister" data-gathering 
 agencies. So far, the Bureau has not been directly identified as a 
 culprit. But some articles and many public statements leave the 
 distinct impression that it has that potential. This implication all 
 too often occurs because the potential danger is viewed in 
 isolation. Current Bureau procedures for confidentiality are never 
 mentioned. Bureau history as to confidentiality of data is not 
 considered. The "watch-dog" nature of the Congress is not raised. 
 Many more aspects of the problem must be considered before one 
 can draw conclusions concerning the abusive or potentially 
 abusive nature of a data-collection system. And I will attempt to 
 deal with these later in this talk. 
 
 But, statements by elected officials, and headlines like 
 those I have mentioned, make it apparent that regardless of the 
 actual situation there is a clear public concern that all is not as it 
 should be. The public reads reports of government officials 
 abusing data collected about individuals. The Census Bureau is a 
 collector of information about individuals. The public is con- 
 cerned that the Bureau should not abuse such information. Part 
 of the public may already believe that such information is abused. 
 Regardless of how one wants to categorize public feelings 
 concerning data collection by the government no one would deny 
 the great and growing anxiety that there is abuse of data— that 
 people either as individuals or in the mass are being manipulated. 
 
 Anxiety 
 
 I see two types of anxiety reflected in these headlines. One 
 is on the part of the individual, the other by society. 
 
 Let's look briefly at society. If one were to generalize about 
 the "data-collection" concerns of society or large groups of 
 individuals, this concern might be broken into two parts. 
 
 First, from the myriad of advertising and mass media cam- 
 paigns, the fear arises that unconsciously the public is coerced to 
 act in ways actually opposite of the individual's best interest. 
 That is, that stimulants are being applied in a scientific and 
 ordered fashion to cause individuals to behave according to 
 
 23 
 
THE RIGHT OF PRIVACY AND THE NEED TO KNOW 
 
 the wishes of the anonymous "they" of corporate America and 
 Madison Avenue. 
 
 Much of this concern has arisen since, and is no doubt 
 linked to, the birth and growth of television and computers and 
 the fact that they have become part of everyday America. 
 
 It has even been alleged that political candidates are being 
 disguised in order to "sell" them to the public— a public that 
 would have rejected them out-of-hand had it not been for the 
 so-called magic of electronics. 
 
 By the late 1950's, computers and television had become 
 part of the everyday scene in America. Vance Packard's book The 
 Hidden Persuaders crystallized a fear that had been forming 
 for some years— that increased knowledge about society would 
 lead to the ability to manipulate the public mind. Words such as 
 "brainwashing" and "subliminal advertising" were current on the 
 cocktail party circuit. 
 
 Raymond Bauer, writing in the Harvard Business Review 
 a few years later, put this fear into perspective. He said that the 
 fear of manipulation is an old one on the part of society. He 
 pointed out that superstitions such as possession by demons 
 and belief in witchcraft have been replaced by fears of the 
 unknown potential of new technological inventions. 
 
 But, Bauer went on to state that those who desire to 
 manipulate the public mind are always one jump behind the 
 public mind. Bombarded by television, "education," and the 
 experience of living in a technological society, you and I become 
 more sophisticated, and our resistance to persuasion increases. A 
 case in point is the consumer movement. Who could have 
 foreseen such a powerful grassroots phenomenon just a few years 
 ago? 
 
 The second concern is more subtle and further from reality 
 than advertising's efforts to control a sophisticated public. 
 
 In Imperial Rome the politicians discovered that games and 
 circuses plus free grain distracted the masses. Therefore, no 
 formidable public opinion ever arose on a variety of extremely 
 important but remote issues such as land reform or the decline of 
 the senatorial system. 
 
 Today, fears have arisen that aggregate data theoretically 
 compiled to show social needs, desires, and minimum levels of 
 fulfillment will allow manipulators to promise what the people 
 want to hear while providing the minimum desired— the "games 
 and circuses" for the people. In other words, by using statistical 
 data the politician is able to pacify the voter while doing his own 
 thing for his "special interests." 
 
 This sort of argument ignores a very fundamental truth 
 about the American Government's statistical programs. That is, 
 simply, that every possible effort is made to produce low cost 
 statistics, available to all groups within our society, so they may 
 analyze and use them in decisionmaking. Everyone admits that 
 statistical information used properly is one of the best tools 
 available for decisionmaking. 
 
 At a recent conference of German and Swedish scholars in 
 Stockholm, it was suggested that one of the finest defenses 
 against the possible manipulation of society was the use by 
 interested groups of the same statistics to influence the manipu- 
 lators. Some of the more successful actions of consumer advocate 
 Ralph Nader certainly serve as examples of putting this advice to 
 good use. So long as no one group holds a monopoly on the 
 statistical information produced by the Bureau and other agencies 
 I cannot imagine the people of our country becoming mere 
 "pushovers." 
 
 Now let's look at the individual's personal concerns. At the 
 top of the list is the question of the individual dossier. The 
 potential abuse of information collected for various administra- 
 tive purposes exists when combining it, without the individual's 
 knowledge, into a hidden and unalterable record of past 
 performance, to be used possibly for questions of job tenure, 
 promotion, or financial loan suitability. 
 
 If there is an area which requires our attention this is it! We 
 must prevent information collected for one purpose from being 
 used later in a way which will injure either the individual or his 
 family. In 1973 the Department of Health, Education, and 
 Welfare released a widely acclaimed report dealing squarely with 
 the above problem and recommended that there be no adminis- 
 trative record kept of which the data subject was not aware, nor 
 could data collected be used for any unstated purpose without 
 the permission of the data subject. At the moment there are 
 several bills before the Congress that would legalize some of the 
 HEW recommendations. 
 
 And this is a problem that we all face; there is enough truth 
 in the average man's fear of manipulation that the fear is hard to 
 allay. But, almost without exception, the abuses discussed 
 concern the use of individual data collected for administrative 
 purposes and action against specif ic individuals, such as tax records. 
 
 Also on the list of the individual's personal concerns is the 
 "mailing list." Many persons are annoyed by the presence of their 
 names on mailing lists and consider this method of marketing an 
 invasion of their privacy. 
 
 I sometimes play a game when I order something by mail. 
 My middle initial is "P," but I often put down another letter to 
 determine if the company I buy something from sells or uses my 
 name for another purpose. Actually, I don't mind being advised 
 by mail of services and products which I might be able to use. I 
 just want to know how somebody got my name. 
 
 The fear of the misuse of personal information is further 
 exaggerated by the popular image of the computer. Spy movies 
 and television have fostered a public overestimation of the 
 abilities and applications of the computer. Together with recent 
 headlines, this has led to fears of an ominous national data bank 
 which would store every facet of our personal lives for instant 
 retrieval by any government agency. 
 
 Let me state that for this fear to be realized— for us to move 
 that close to George Orwell's nightmare-the United States would 
 have to abrogate not only current law but its entire democratic 
 tradition. Let me emphasize this by quoting a key portion of a 
 paper written by Otis Dudley Duncan which concerned plans for 
 the 1970 census. 
 
 . . .in this country we have proved that a statistical 
 system can incorporate rigid safeguards of confidentiality. 
 The institutionalization of these safeguards has proceeded 
 to the point where it is inconceivable that they would break 
 down, except in the catastrophic event of a breakdown in 
 our whole system of institutions protecting the rights of the 
 individual. In the case of such a catastrophe, my guess is 
 that much more direct ways of infringing these rights would 
 be found than that of making inappropriate use of 
 statistical records secured ostensibly in confidence. 
 
 Nevertheless, the public conception of the computer is 
 often that of a villain, because of either its potential for abuse of 
 personal information, or irritating arguments with department 
 store computers about incorrect billing. 
 
 24 
 
THE RIGHT OF PRIVACY AND THE NEED TO KNOW 
 
 Yet, I think the computer has made a profound contribu- 
 tion to the public good— and done it so well we take it for 
 granted. One example is the millions of checks the Social Security 
 Administration issues each month. Those checks simply wouldn't 
 arrive as fast or as accurately with manual processing. Further- 
 more, as regards our topic of privacy, the advent of the computer 
 has done a great deal to protect an individual's privacy, as I hope 
 to demonstrate a little later in this talk. 
 
 I have no objections to the press or elected officials 
 investigating the potential for misuse of any system containing 
 information on our citizens. However, I am very concerned that, 
 as real culprits are identified, the Census Bureau will be tarred 
 with the same brush. Yet, there has been no real investigation of 
 Bureau procedures, its standards of privacy and confidentiality, 
 or the importance of the societal need that the Bureau's products 
 fill. People are largely unaware the Census Bureau is the only 
 government or private agency that collects data on every 
 household in America. Nor, does the public at large realize what 
 the Bureau does to ensure the privacy and confidentiality of the 
 respondents. 
 
 Because of the concerns I've expressed, I would like to 
 detail the Bureau's philosophy on the citizen's privacy and 
 confidentiality of the information it collects. 
 
 Logic of Information 
 
 We, at the Census Bureau, believe our mission is to gather 
 accurate, timely, and complete data from individuals, businesses, 
 and governments, with a minimum of intrusion into their privacy, 
 and to make publicly available general, anonymous statistical 
 summaries of that data, specifically designed to respect the 
 confidentiality promised each respondent. 
 
 First, we should address the question of whether we need a 
 Census Bureau at all. What would happen if the Bureau closed 
 down or greatly restricted its collection of data? Without census 
 statistics much of the planning in the public and private sectors 
 would have to be based on general assumptions which would not 
 be verified. 
 
 You may ask what sort of uses are census statistics put to? 
 Bureau-generated figures include those used to determine the 
 unemployment rate. Federal revenue sharing allocations, the 
 gross national product, characteristics concerning the low income 
 population and minorities within our country— including figures 
 on birth, education, and unemployment among these subgroups- 
 projections of population, birth expectations, changing age 
 patterns of the population, the balance of U.S. imports and 
 exports, and crime victimization characteristics by socioeconomic 
 status, to name but a few. I should add that all these examples do 
 not include the many uses of the decennial census. 
 
 My examples are intended to emphasize the statement that 
 without accurate and timely census statistics decisionmaking both 
 inside and outside the government would be handicapped. 
 
 It is probably obvious to you, and I would be less than 
 candid if I did not say, there is an inherent conflict in gathering 
 data from individuals. That conflict is between the amount of 
 privilege granted the individual to decide whether to disclose 
 information about himself and society's use of mandatory 
 processes to obtain the information it needs for valid purposes. 
 
 I have not mentioned the individual's right to be let alone 
 because I am not sure that, by itself, such a right exists. The 
 choice of individuals to live in a group, clan, or society implies a 
 
 certain obligation to that group. If an individual obtains benefits 
 from such a living situation there is little doubt he will be 
 expected to contribute to the advancement or continuation of 
 that group. Few groups can be identified where an individual is 
 allowed to reap the benefits of society and yet does not con- 
 tribute in some way. The obligations of dwelling in a highly 
 complex industrial civilization are perhaps even greater than ever. 
 There is little doubt that the individual as a single human unit 
 derives benefits from dwelling among his fellow men. 
 
 For instance, no one forces a person to apply for a driver's 
 license, to register to vote, to obtain a passport, or to practice a 
 licensed profession. But if a person wants to do any of these 
 things, he or she must provide certain information which society 
 has deemed necessary to establish qualifications. These are basic 
 steps of order that society takes to protect itself. 
 
 What, then, should be the price charged an individual for 
 the services and benefits, both realized and potential, he receives? 
 I would suggest a broad definition which would require each 
 person to cooperate as fully with those persons who represent the 
 institutions of society as is necessary for the smooth functioning 
 of the society. Society, in return, must meet most of its 
 responsibilities to the individual's needs, most of the time. To do 
 this, those officials and decisionmakers responsible for the 
 orderly operation of the socioeconomic conglomeration we call 
 "the state" must have accurate information. They must get this 
 information from the individuals of the society to determine and 
 evaluate needs, demands, trends, and a variety of facts indispen- 
 sible to the intelligent running of such a system. 
 
 Basic to this discussion then is the question: What is the 
 right of privacy? Is that right something that can be brushed aside 
 anytime a governmental organization believes it needs a new 
 brand of information to assist the individuals? The right to 
 privacy is indeed a very easy term to evoke, but a very difficult 
 one to define. 
 
 American legal and academic scholars have wrestled with 
 the problem of privacy ever since 1890 when Samuel Warren and 
 Louis D. Brandeis, later a famous Justice of the Supreme Court, 
 published in the Harvard Law Review their well-known article 
 entitled "The Right To Privacy," which they defined as a right to 
 be let alone. 
 
 Warren and Brandeis developed their quasi-legal definition 
 of privacy by viewing it as a "right" to be secure from undue 
 interference with one's person, papers, property, or thoughts, 
 having traced its development through the history of common 
 law. 
 
 Today, several Western European nations have commissions 
 reviewing existing privacy safeguards and planning additional 
 measures in order to protect the individual's desire and need to be 
 let alone. I have already mentioned the proposed American 
 legislation intended to deal with this matter. The Swedish 
 legislature has recently enacted a data law designed to cope with 
 problems of individual privacy. Some German states have for 
 several years employed a data privacy and confidentiality 
 ombudsman whose task is to act as an independent investigator, 
 reporting directly to the elected officials. 
 
 More recently, German psychologists have attempted to 
 define privacy as those areas of individuals' lives in which they 
 can act without having to fear that information may be passed on 
 in a way dysfunctional to themselves. Thus, Muller and Kuhlmann 
 have added to the privacy concept the idea of selective 
 transmission of information, be it to family, friends, one's doctor, 
 or a government organization. They view privacy in the modern 
 
 25 
 
THE RIGHT OF PRIVACY AND THE NEED TO KNOW 
 
 sense as including the selective transmission of information about 
 one's self. 
 
 For the purposes of our discussion let's accept the notion 
 that the word "privacy" refers to a sphere in which the individual 
 is able to decide what will be transmitted to others. In addition, 
 this sphere also includes certain confidential relationships where 
 the individual believes that information transmitted, either by 
 choice or compulsion, will go no further than the parties to that 
 relation, and will only be used for the purposes that were identi- 
 fied at the time the information was imparted. Given this notion 
 we can see that the term "privacy" does not mean a sacrosanct 
 area where no questions can be asked; that is only part of the 
 definition. On top of that we have a larger area where questions 
 may be asked on a voluntary or mandatory basis, but the indivi- 
 dual understands and believes the answers he gives will not be 
 used for other than what he has been told or perceives. 
 
 The respondent's "security" of belief, based on the promise 
 of the receiver, stated or implied, that information will only be 
 used in specified ways, creates a bond which I would call a 
 "confidential relationship." To me, confidentiality must be 
 considered when anyone attempts to assess or balance individual 
 privacy and society's need to know. But, what exactly are 
 confidential relationships? 
 
 Census confidentiality relates to individual information in 
 the hands of institutions. It is important to reinforce the 
 distinction between personal information gathered for statistical 
 purposes as opposed to that collected for administrative tasks. 
 The information could be the same in both cases, but administra- 
 tive records are intended to affect the individual directly, for 
 instance, those used by the Internal Revenue Service or the Social 
 Security system. Statistical records, such as those maintained by 
 the Census Bureau, do not directly affect the individual. 
 
 All that I've really been trying to point out is that 
 "privacy" doesn't exist in an absolute sense anymore than 
 freedom does. There is no doubt that our constitution is less than 
 specific about all the aspects of individual privacy and, yet, our 
 courts have been able to find several guideposts for dealing with 
 specific cases. As our Supreme Court is called upon to adjudicate 
 constitutional questions pertaining to the right to privacy, we 
 must remember that privacy, like freedom, has meaning only in 
 the context of human society, and our society is rapidly 
 changing. The crux of the issue is this: As American society 
 becomes more complex, we need to know more about ourselves 
 in order to establish priorities and properly allocate our human, 
 financial, and natural resources. 
 
 "Real World" Decisions 
 
 All of this merely represents our philosophical approach to 
 matters of individual privacy and respondent-Bureau confiden- 
 tiality. But, none of this goes very far in showing you how the 
 Bureau faces the tough "real world" decisions constantly con- 
 fronting it. 
 
 For example, in the rural style of life that existed in 1790, 
 it would have been hard to justify the government's interest in 
 whether or not a household had its own bathroom facilities. But, 
 today, with society's commitment to eliminate slums and sub- 
 standard housing, information about sanitary facilities is needed 
 to accurately pinpoint the number and general location of such 
 housing so the taxpayer's dollar is spent wisely. 
 
 I've cited to you the question about plumbing because it is 
 one of the many well-publicized examples that critics of the 
 
 26 
 
 Bureau use to show how meaningless, silly, personal, or unneces- 
 sary our questions are. To claim that this census question now 
 constitutes a violation of an individual's privacy strikes me as a 
 rebirth of the "know nothing" philosophy of the 19th century. 
 Put another way it is an argument that the interest of the 
 individual refusing to answer such a question transcends the 
 interest of the society and has priority over public efforts to 
 eliminate slums and promote accurate resource allocation. 
 
 But the question of a household's plumbing is not the only 
 example. I don't need to tell this audience that the salaries and 
 pensions of millions of Americans are tied to variations in the 
 cost of living, which is gauged by the Consumer Price Index. 
 However, recently appearing before a committee of the House of 
 Representatives, I was questioned intensely about the census 
 questionnaire which forms part of the Consumer Expenditure 
 Survey, which is the basis for the Consumer Price Index. One 
 member wanted to know why the entry "hair and scalp care for 
 men and boys" was important. Out of context, I admit it sounds 
 ludicrous, but as one component of a family's total spending 
 record, it is important. 
 
 The Consumer Expenditure Survey was taken of a scientifi- 
 cally selected sample yielding 17,000 household interviews. As 
 with all the Bureau's current household surveys, participation was 
 voluntary. I should add that despite the fact that the survey was 
 very lengthy and required that the interviewer visit the respond- 
 ent several times over a 2-year period the Bureau obtained 
 approximately a 90-percent response rate, indicating excellent 
 cooperation. And, the answers of each family are very important, 
 since the 17,000 households represent the entire civilian non- 
 institutional population of the United States. 
 
 As long as responsible Government leaders determine that a 
 detailed picture of family spending is essential, there is no other 
 way to obtain it than to ask such detailed questions. 
 
 Another example is the census question about how a person 
 gets to work. In 1970, some people felt this was none of the 
 Government's business. But when the energy crisis hit with full 
 force, the only complete statistics about the Nation's commuting 
 habits came from the total of each person's answer to that 
 question. The Bureau was able to respond with statistics about 
 the number of drivers and passengers for each metropolitan 
 area— figures which showed that a total of 26 million persons 
 drove to work alone each day. 
 
 I've given you a few examples of questions that some 
 people feel are an unnecessary invasion of an individual's private 
 sphere and I've tried— briefly— to explain why they are relevant 
 and necessary. 
 
 I think that Senator Sam Ervin in his role as Chairman of 
 the Subcommittee on Constitutional Rights summed the problem 
 up well when he said: 
 
 Somewhere a balance must be struck between the 
 individual's desire to keep silent and the Government's 
 needs for information. If it is proved necessary to invade 
 certain rights, clearly it is the constitutional duty of 
 Congress to establish precisely how and under what 
 circumstances this may be done. 
 
 Congress has been doing exactly that for almost a century. 
 Since the act which provided for the 1880 census, the laws 
 protecting the confidentiality of the information given in 
 response to census questions have been progressively strength- 
 ened. While, at the same time, what the Bureau can collect has 
 been defined as well. 
 
THE RIGHT OF PRIVACY AND THE NEED TO KNOW 
 
 This was not always the case. To begin with, until this 
 century there was no permanent census organization. Federal 
 marshals supervised the enumeration from the first census in 
 1790 through the count of 1870. In fact, the rules for the first 
 five censuses required that the marshals post copies of the list of 
 names in two public places in their districts. 
 
 Between 1840 and 1870 there were no legal restrictions, 
 but census takers in the field were instructed to treat all 
 information they gathered as strictly confidential. I don't know if 
 the urge to gossip was stronger than those instructions, but 
 beginning in 1880 the law required all enumerators to take an 
 oath not to disclose any personal information. Oddly, this 
 requirement did not extend to their supervisors. That loophole 
 was closed by 1900, when all census employees were made 
 subject to a $500 fine for violation of their oath. 
 
 In 1902, the permanent Census Bureau was established. It 
 had become apparent the need for statistics was a continuing one. 
 Also, the size of the job was such that a full-time, trained staff 
 was much more efficient than setting up and disbanding a new 
 organization every decade. 
 
 Up until 1910, census law required the Director to furnish 
 on demand to Governors or heads of municipal governments 
 certain parts of an individual's return— the name, age, sex, birth 
 place, and race. The act for the 1910 census changed that 
 wording to read that the Director could, at his discretion, furnish 
 information for genealogical and other proper purposes. 
 
 For the Bureau, 1910 also marked the start of another 
 tradition— the presidential proclamation. The one issued by 
 President Taft told the American people their replies to census 
 questions were to be used only to compile general statistical 
 information, and that their answers were protected by law. In 
 part, it read: 
 
 The census has nothing to do with taxation, with army or jury 
 service. . .or with the enforcement of any national. State, or 
 local law or ordinance, nor can any person be harmed in any 
 way by furnishing the information required. 
 
 The current law under which the Census Bureau operates is 
 Title 13 of the U.S. Code, most of which dates from 1929. This 
 law is very specific when it comes to personal information. It 
 requires that information obtained from an individual be used 
 only for statistical purposes. It also requires that published data 
 be in such a form that it is not possible to identify an individual 
 or a single business establishment. The law stipulates that no one 
 other than sworn officers and employees may have access to 
 individual information, and each census employee has signed an 
 affidavit of nondisclosure to uphold the law. 
 
 The current law still allows the Director at his discretion to 
 provide copies of individual information for genealogical and 
 other proper purposes. The key word here is "discretion." Over 
 the years, the application of this rule has become restrictive 
 rather than permissive. 
 
 In current Bureau practice, the term "confidentiality" 
 represents nothing less than a clear extension of an individual's 
 right to privacy. I think the best way of showing this is to review 
 the Bureau's track record regarding the confidentiality of 
 individual data. 
 
 Track Record 
 
 Until recently, most of the Bureau staff assumed that 
 
 individual information collected had always been held in strictest 
 confidence. I must report that this has not always been the case. 
 But looking at the way we did things in the past, and comparing 
 them with today's practices, I am even more certain that our 
 current position is a very strong one. 
 
 Between 1900 and the mid-1 920's, there were authorized 
 releases of individual data which were then considered proper; 
 today they would cause a storm of protest in the press, in the 
 courts, and in Congress. As far as we know this practice caused no 
 such outcry then, although complete records do not exist. 
 
 We do have some information on one case which demon- 
 strates the type of situation in which it was considered proper in 
 the past to release data about individuals. This release occurred in 
 1918, during World War I. Congress had passed a war powers act, 
 presumably the basis for an extreme use of census data. 
 Information about individuals was given to the Department of 
 Justice for use as evidence in prosecuting young men who claimed 
 they were too young to register for the draft. While we do not 
 know the exact circumstances surrounding the release, we do 
 know that personal information for several hundred young men 
 was released to courts, draft boards, and the Justice Department. 
 
 The Bureau stopped such releases during the 1920's, a 
 position which was made official in 1930 by an opinion from the 
 Attorney General. His opinion said that even the name and 
 address of an individual was confidential. 
 
 At about this same time, the Secretary of State asked for 
 data about individual farms in Stevens County, Wash. Clouds of 
 sulphur dioxide from a smelter located across the border in 
 Canada had caused extensive damage to crops in the United 
 States, and the matter had been handed over to an international 
 tribunal. The Census Bureau refused to release the information, 
 and the tribunal decided not to press the point, even though it 
 felt it had the necessary power. The reason? It would have caused 
 the U.S. Government to breach a promise it had made to its 
 citizens as well as to violate the U.S. law of census confiden- 
 tiality. The Census Bureau did provide aggregated county data 
 which allowed the court to award damages to the Washington 
 farmers. 
 
 Now we jump to 1941. It's hard to imagine now but, with 
 World War II underway, there was near hysteria about the 
 Japanese-Americans living on the West Coast which led to one of 
 the most embarrassing moments in U.S. history, the internment 
 of large numbers of these loyal Americans. At the height of this 
 feeling, the Secretary of War requested the Census Bureau supply 
 the names, addresses, and ages of all persons of Japanese 
 extraction living on the West Coast. This time, in spite of the 
 national emergency, the Bureau held to its position on confiden- 
 tiality of individual records and refused. However, the Bureau did 
 supply summary data at tract level, but not individual data. I 
 should add that today tract level data is part of the regular 
 publication program of the Bureau. But, as in 1941, no individual 
 disclosures, even by implication, are allowed. 
 
 In 1947, during the rising concern about possible Commu- 
 nist infiltration and sabotage, the Attorney General requested 
 information from census records about certain individuals for use 
 by the FBI. Again, the request was denied. 
 
 A loophole in the law turned up in a case in the early 
 1960's when the courts ruled that file copies of census forms not 
 kept by the Bureau could be subpoenaed. This resulted in 
 Congress amending the law to extend census confidentiality to 
 include even copies of census questionnaires which are kept by 
 businesses for their own files. 
 
 27 
 
THE RIGHT OF PRIVACY AND THE NEED TO KNOW 
 
 There's a technological aspect of the Bureau's history that 
 I'd like to mention as well. Herman Hollerith, an engineer who 
 had worked as a special agent in the 1880 census, became 
 interested in the problem of how the volume of material gathered 
 in the census was to be processed. After some experimentation he 
 invented the punched card system for recording and tabulating 
 the census returns. Even in its earliest stage of development, the 
 Hollerith system speeded tabulations to such an extent that its 
 merits were demonstrable before the 1890 enumeration began. 
 
 The advent of this new system had dual implications for the 
 question of confidentiality. On the one hand, it removed the 
 actual census return one step farther from human statistical 
 processing. On the other hand, it made possible the collection of 
 even more information on individuals. It is probable that the 
 Hollerith system enhanced the anonymity, and thus the confiden- 
 tiality, of census data. Technologically it was the forerunner of 
 modern, computerized recordkeeping, which can provide con- 
 siderable confidentiality for respondents. 
 
 Briefly, that is a summary of how confidentiality grew to 
 be an integral part of census taking. Keeping that information in 
 mind, and my earlier remarks about the generalized fear of the 
 computer, let's look at how a modern census is taken and 
 processed. 
 
 Modern Census Taking 
 
 As you know, we have increasingly utilized the U.S. postal 
 system. In 1960, forms were mailed to each household and were 
 picked up by enumerators. For 1970, the forms were again 
 mailed to all households. But those people living in larger 
 metropolitan areas returned their forms by mail. This meant that 
 in about three-quarters of the Nation's homes personal informa- 
 tion did not have to be given face-to-face to a neighbor or a 
 stranger. In Great Britain, a 1973 parliamentary paper dealing 
 with the security of the census of population indicated that the 
 Council of the Royal Statistical Society viewed as a problem 
 having to give personal information to someone who knew the 
 respondent. 
 
 After all the forms are collected the data on them must be 
 transferred to the computer for tabulation. It used to be that the 
 data on each form was manually transferred to punchcards, and 
 the punchcards fed to computer tape. Now, we bypass this 
 laborious process. The forms are microfilmed and returned to 
 storage. This is the last time each original form is handled until it 
 is destroyed. 
 
 The first page of the census form is not microfilmed. This 
 page has the address of the household. So the rolls of microfilm, 
 which have names and personal information, contain only a 
 geographic code relating that information to the block on which 
 the household is located, and the address register is maintained in 
 a separate file. 
 
 Another sophisticated machine transfers the microfilmed 
 information onto computer tape. 
 
 Even this is not enough to guarantee that a person could 
 not be identified in the statistical summaries. Some areas have 
 such a small population that it would be possible, by deduction, 
 to know whose characteristics are in the tables. Our computer 
 program is set up to suppress any characteristics that would 
 enable an analyst to identify individuals. So, as I've said a couple 
 of times already, you can see how the computer is a direct asset 
 to contributing to the confidentiality of the information col- 
 
 28 
 
 lected. The Bureau is able to accurately and rapidly produce 
 detailed statistics, not possible if the editing of potential 
 disclosures had to be done by hand. 
 
 We are examining other techniques for protecting confiden- 
 tiality. These include rounding numbers to the nearest 5, and a 
 "random noise" system, in which values of one and negative one 
 are scattered throughout the tabulations, balancing to zero at 
 certain geographic levels. Such a system would have no substan- 
 tial effect on statistical analysis. 
 
 When it comes to suppression of data from the economic 
 censuses, even the cutoff points are confidential— because that 
 information by itself could be used for deduction, since the 
 numbers involved are so much smaller than population figures. 
 
 I also should mention here the Public Use Sample estab- 
 lished in 1960. These are not summaries, but individual census 
 records minus certain data which ensure that the individual 
 cannot be identified. The smallest area identifier for which these 
 records are available is 250,000, and even then certain data have 
 been truncated to avoid identification. An example would be 
 extremely high salary figures, for instance, everything over 
 $50,000 is simply marked "$50,000 and over." 
 
 These samples have proven of great value to the academic 
 and business communities for research and for determining if 
 special tabulations would provide the summary data desired. 
 
 When the tabulation of a census is finished, the original 
 paper forms which have been stored in guarded buildings on a 
 Government facility are destroyed. They are shipped in sealed 
 boxcars and recycled, with Bureau officials watching until they 
 drop into the pulping vats. 
 
 That leaves the microfilm. Where does it go after we are 
 finished processing the data? The rolls are sent to the Personal 
 Census Service Branch in Pittsburg, Kans., which we commonly 
 refer to as the Age Search Service. This is a unique self-supporting 
 operation which has helped millions of people. Every day the 
 Bureau receives about 1,300 requests from people who need to 
 verify some item of information about themselves. Most are for 
 substitutes for birth certificates which either never existed or 
 have been lost or destroyed. People need them to qualify for 
 retirement, for social security, for Medicare, to get a passport, 
 and for many other uses. For a very small fee the Age Search 
 Service will search old census records and issue a certificate which 
 has legal standing. 
 
 This service is provided only at the request ot the person 
 himself. For example, a son cannot ask about his father unless he 
 has a power of attorney or a death certificate. This operation is 
 the only use made today of the Director's authority to release 
 personal information at his discretion. 
 
 Finding information for those who request it is not an easy 
 job. It takes an expert to utilize the microfilm. Since the census is 
 based on addresses, not names, there is no such thing as a master 
 list of records arranged alphabetically by name. For the correct 
 reel of film to be located, the person making the request must 
 supply information about where he lived. 
 
 Remember, I said the microfilm for 1970 does not contain 
 both name and address. In order to work backwards and relate 
 name with address, it is necessary to have access to the geographic 
 coding information on the film and the master list of addresses. 
 
 The very size of the U.S. population helps to guarantee 
 confidentiality. It took some 5,000 miles of microfilm to process 
 the 1970 census. For us to make this process of working 
 backward any easier would be extremely costly and would, in 
 theory, weaken the protection of confidentiality. 
 
THE RIGHT OF PRIVACY AND THE NEED TO KNOW 
 
 Now, where does the microfilm of past censuses exist? The 
 records of the counts from 1790 through 1880 are accessible to 
 the public in the National Archives. Data in these enumerations 
 were not gathered under laws of confidentiality. The census of 
 1890 is almost nonexistent, having been mostly destroyed by 
 fire. 
 
 The Pittsburg, Kans., unit has microfilm for 1900 through 
 1960 and late this year will have the 1970 records set up to be 
 able to answer the queries which have been coming in. A copy of 
 the 1960 and 1970 records will be held under security conditions 
 at Jeffersonville, Ind., so that if one copy were destroyed the 
 whole census would not be lost. Incidently, a tornado lifted one 
 corner of the roof of the Pittsburg building a few years ago. 
 
 Copies of the 1900 through 1950 records are also held by 
 the National Archives. These were sent to the Archives for 
 storage. However, in December of last year, the Archives opened 
 the 1900 census to limited access by qualified researchers, a move 
 opposed by the Census Bureau. While this access is under 
 controlled circumstances, we feel at the very least it violates in 
 principle the rights of the estimated seven million persons still 
 alive who were counted in the 1900 census. 
 
 The law under which the Archives operates says Govern- 
 ment records shall be made public after 50 years, unless an 
 interagency agreement stipulates a longer period of time. In 1952, 
 the Director of the Census Bureau and the Archivist agreed that 
 census records should remain closed for 72 years— an average 
 person's lifetime. The Bureau's position is that the 1952 
 agreement was in excess of the Director's authority and, 
 therefore, is invalid. However, an opinion from the Justice 
 Department did not support our view. 
 
 Remember that there was a difference in the status of 
 protection between the census of 1900 and that of 1910. The 
 1910 count was preceded by a Presidential proclamation-a 
 promise from President Taft which stated: "There need be no 
 fear that any disclosure will be made regarding any individual per- 
 son or his affairs." 
 
 The obvious question is how long does that promise and 
 the law's guarantee of confidentiality apply? A lifetime? One 
 hundred years? Or forever? The Bureau hopes Congress will close 
 this final loophole in the laws relating to census confidentiality. 
 
 Congress at the moment has a lot to consider in the area of 
 privacy. Some 100 bills are pending in the House and Senate. The 
 basic question seems to be not whether something must be done 
 to ensure privacy and protect it, but what, and by whom. Four of 
 those bills deal with census information. Eight of them would 
 establish a Federal Privacy Board or some committee or com- 
 mission as an overall authority. Many of the bills would allow the 
 citizen the right to inspect his own records, correct them, and 
 bring suit for damages resulting from incorrect or misused 
 records. 
 
 Much of the "data bank" fear centers on the use of a 
 person's Social Security number as a universal identification 
 number. One bill would expressly ban use of Social Security 
 numbers for any identification other than the retirement system. 
 The armed forces scrapped the old ID serial number several years 
 ago in favor of Social Security numbers, and several States are 
 using these numbers on their drivers' licenses. It is interesting to 
 note that Social Security numbers have not been asked for in the 
 decennial census, where answers are mandatory. 
 
 We have asked for these numbers in some voluntary 
 surveys. This allows us to match current answers with informa- 
 tion the Government has already gathered, to evaluate the 
 
 accuracy and consistency of the statistics we produce. 
 
 One of the most common complaints from the public is 
 that several agencies of the government ask the same questions. 
 So, where it is feasible, and when authorized by law, the Census 
 Bureau utilizes the records of other agencies. It does this, even 
 though it has the authority to go out and collect the data itself, 
 to save money and to lessen the burden on the public of 
 answering questionnaires. When we do use such administrative 
 records, the information comes under the same rules of confiden- 
 tiality as if we gathered it ourselves. 
 
 While we're on the subject of the burden to the public, let 
 me tell you that the Bureau spends a lot of time and effort in the 
 research and development of questionnaires and techniques of 
 taking surveys and censuses. 
 
 For a question to be used, its propriety has to be weighed 
 against the social need for the information. The wording is also 
 very important. A question must result in the maximum of 
 information for society, while causing a minimum of inconven- 
 ience for the individual. 
 
 Here again, the questions change as society changes. When 
 the Nation was not crowded— when big families were needed to 
 run farms— a question asking how many children a couple 
 expected to have would have been out of line. With today's 
 concern about the size of the population, such information 
 becomes very important. 
 
 The concern for the right to privacy was underlined when 
 President Nixon established a blue ribbon panel to review both 
 government and industry practices related to the collection and 
 use of personal information. 
 
 At the same time. Commerce Secretary Frederick Dent has 
 directed the National Bureau of Standards to find out how 
 computer information may be made more secure. 
 
 And at the Census Bureau, we have established an Advisory 
 Committee on Privacy and Confidentiality, staffed with experts 
 from the entire spectrum of data gatherers and users, to advise us 
 on procedures to maintain and strengthen our record of confiden- 
 tiality. 
 
 You might well ask why we, at the Census Bureau, are so 
 concerned, if our record is good and our intentions are clear. 
 Aside from the moral implications, I'll give you a very practical 
 answer. A census or a survey is only as good as the contract or 
 trust with the people who answer the questions. If the public 
 feels we are not keeping our word that their answers will be kept 
 confidential, or even that the potential for such violation of their 
 trust exists, their answers will not be as accurate or given as 
 willingly. If this occurs on a large scale, the quality of the 
 summary statistics will deteriorate. And if this occurs, the Nation 
 has lost a prime decisionmaking tool, and society will be the 
 loser. 
 
 This would be a tragedy! It would come just as more and 
 more decisionmakers in the public and private sectors are 
 becoming aware of how valuable census data is to them. And it 
 would come as the Bureau's main thrust is to increase the utility 
 of the data it gathers. 
 
 To the Census Bureau, a promise is a promise. The calendar 
 may read 1984 in 10 years, but I want to make it clear that as far 
 as the Bureau is concerned, 1984 will never come. 
 
 In his remarks to the most recent meeting of the Advisory 
 Committee on Privacy and Confidentiality, Secretary Dent said: 
 "I think perhaps the strongest brand name in America might be 
 that of the Census Bureau." 
 
 We plan to keep it that way. 
 
 29 
 
MONITORING THE NATION'S HEALTH 
 
 Edward B. Perrin, Director, National Center for Health Statistics; 
 Douglas Williams, Chief, Data Use Laboratory, National Center 
 for Health Statistics 
 
 "Health care is an industry on which there is relatively little information available. 
 Decisions are being made daily in Washington and Sacramento and other places around 
 the country which involve millions and billions of dollars based on very little information, 
 and suddenly it has occurred to the people in positions of authority that there should be 
 more information and that we should have it now." 
 
 Edward Perrin 
 
 Dr. Perrin: It's a pleasure to be here today and to have Doug 
 Williams with me. 
 
 Before I begin our presentation I think a few general 
 comments are appropriate. By the standards of the Federal 
 bureaucracy the National Center for Health Statistics is a fairly 
 small organization, with a budget of approximately $18 million. 
 It is historically, of course, much younger than the Bureau of the 
 Census and, in fact, an offshoot of the Bureau, though some 
 further background will help you better appreciate that. 
 
 When the Public Health Service was established, about 15 
 or 20 years ago, it was decided that the vital statistics system of 
 the United States should be managed by that organization for the 
 Federal level. This formed the basis of what was to become the 
 National Center for Health Statistics. 
 
 Then, about 10 or 12 years ago, Congress passed a very 
 innovative law introducing the National Health Survey, a mecha- 
 nism for gaining information about the Nation's health. 
 
 The National Health Survey and the vital statistics system 
 were merged into the organization called the National Center for 
 Health Statistics which has existed for approximately 12 years. 
 In these 12 years the thrust of the organization has changed 
 remarkably because health has changed dramatically. 
 
 Let me tell you where the National Center for Health 
 Statistics exists in the Federal structure. We are in the Health 
 Resources Administration of the Department of Health, Educa- 
 tion, and Welfare. We are under the jurisdiction of the Assistant 
 Secretary for Health, as are the agencies of Food and Drug 
 Administration, the National Institutes of Health, the Center for 
 Disease Control, the Health Services Administration, and the 
 Health Resources Administration. The National Center for Health 
 Statistics is in the Health Resources Administration, together 
 with Health Manpower, Health Planning, and Health Services 
 Research. So, besides being in a different department from the 
 Bureau of the Census, our functions are considerably more 
 limited, in the sense that our basic mission is entirely devoted to 
 gathering data on the health of the Nation's people. 
 
 Now, there are a lot of ways that an organization can exist 
 within the Federal bureaucracy. One way is to be the figment of 
 some administrator's imagination, which was the case with us for 
 15 years. However, that makes one very dependent on the 
 administrator and he can suddenly decide that he doesn't need 
 you any longer. So, our first 15 years were very tenuous. Today, 
 we are established by law and have a separate existence so we are 
 definitely going to be around for sometime to come. Our law, 
 Public Law 93-353, established the National Center for Health 
 Statistics as a Federal data-collection agency. 
 
 30 
 
 Let me just read to you a part of the law which indicates 
 what it is the Congress thinks we should do. Naturally, there's 
 almost always a difference in interpretation between what the 
 executive branch and the Congress think you should be doing. 
 The Congress believes that it is our function to collect statistics 
 concerning (a) the extent and nature of illness and disability of 
 the population of the United States; (b) the impact of illness and 
 disability on the economy of the United States; (c) environ- 
 mental, social, and other health hazards; (d) determinants of 
 health (and that, of course, comes close to the question of health 
 indicators); (e) health resources, including physicians, dentists, 
 nurses, and other professionals, by specialty and type of practice 
 and supply of services; (f) the utilization of health care, including 
 utilization of ambulatory health services, by specialty and type of 
 practice, services of hospitals, extended care facilities, and the 
 like; (g) health care costs and financing, including trends in health 
 care prices and costs, the source of payments for health services 
 and, (h) family formation, growth and dissolution. But it boggles 
 the mind to think that any one organization will ever collect any 
 data on some of these things. The idea is, of course, that health is 
 a rapidly growing industry. The health "industry" cash flow 
 totals $100 billion per year at present. 
 
 Health care is an industry on which there is relatively little 
 information available. Decisions are being made daily in Wash- 
 ington, Sacramento, and other places around our country in- 
 volving billions of dollars based on very little information. 
 Suddenly it has occurred to the people in positions of authority 
 that there should be more information and that we, as a so- 
 ciety, should have it now. We should have it, for example, this 
 week when the National Health Insurance Policy is being 
 hammered out between HEW Secretary Caspar Weinberger and 
 Chairman Wilbur Mills of the House Ways and Means Committee. 
 As is usually the case, people don't really recognize the 
 importance of information until they come to need it and then 
 discover that it's not available and can't be made available in the 
 time that's required. There's definitely a growing awareness of the 
 necessity for good health data at all levels, and we are attempting 
 to fill this need. 
 
 As a data-collection agency, we are interested in a number 
 of very specialized concepts: One is confidentiality of informa- 
 tion which, to us, is extremely important. Here, we follow the 
 lead of the Census Bureau because they have been in business 
 much longer than ourselves and have set a good example. The 
 Census Bureau and our Center are the only two data-collection 
 agencies that, up until very recently, had clauses in the laws under 
 which they operate that allowed them to claim confidentiality for 
 
MONITORING THE NATION'S HEALTH 
 
 the data they collected. Let me say that without our confi- 
 dentiality clause the National Center for Health Statistics would 
 be dead. We operated previously under a law which assigned 
 responsibility to the HEW Secretary. The Secretary then 
 delegated the authority derived from that law to our organiza- 
 tion. The new situation— having our own law— doesn't change our 
 functions, it only changes the fact that we now have a legal 
 existence that separates our mandate from that of the Secretary 
 of HEW. 
 
 The Federal Government has no law which says that it is 
 necessary to collect birth and death data and marriage statistics. 
 The States, on the other hand, have laws which require you to 
 report births. So, the oldest health data system is the vital 
 statistics system. Most of you entered it at birth and continue to 
 deal with it by having to show birth certificates for various 
 purposes, along with the fact that a death certificate must be filed 
 before obtaining a burial permit. That's the way it's enforced in 
 the States. Our center collects these data from each of the 50 
 States. 
 
 The death certificate, which is the backbone of the vital 
 registration system as far as health is concerned, includes some 
 useful health information, demographic characteristics about the 
 deceased, and the cause of death. Now, the cause of death is 
 obviously an important piece of health information. For example, 
 it was the cause-of-death statistics drawn from the vital statistics 
 system that allowed us to draw the inferences about the rise of 
 deaths due to lung cancer, which the epidemiologists associated 
 to smoking. Even so, the system has obvious deficiencies since it 
 is an old system that was set up purely for the administrative 
 purposes of knowing where people were and, as a matter of fact, 
 for getting persons on and off the voting rolls. On the other hand, 
 the system has allowed us to trace, over time, the rise and fall in 
 the prominence of certain causes of death, and we have 
 attempted to infer from these figures something about the causes. 
 
 Clearly, the need for health information goes well beyond 
 those figures that we can obtain from the vital statistics system. 
 For example, the National Health Survey is another important 
 mechanism. This is a survey conducted through the Bureau of the 
 Census. The concept behind the survey is that the vital statistics 
 system does not tell us anything about illnesses that do not lead 
 to death. While mortality tells us several things, morbidity is even 
 more interesting and useful because it will tell us about the load 
 on the provider population, on the economy, and so forth. Our 
 Center, through the Bureau of the Census, selects a probability 
 sample of households throughout the country and conducts an 
 ongoing household survey which asks questions regarding the 
 health of the family during the past 2 weeks. Thus, we have, in 
 addition to the vital statistics system, a household interview 
 survey that asks questions about episodes of illness, episodes of 
 hospitalization, costs of hospitalization, disability, etc. 
 
 A third provider of information is the Health Examination 
 Survey, in which we draw a probability sample of persons around 
 the country, by geographic location. We haul four trailers into the 
 parking lot of a shopping center and connect them together. We 
 convince the people who have been so fortunate, or unfortunate, 
 as to fall into our sample that they should have a physical 
 examination, and conduct physical examinations on a random 
 sample of the population of the United States. We do not ask for 
 volunteers, and our trailers are simply set up in a parking lot for 
 the purpose of easy access. 
 
 Obviously, this is a small sample, approximately 6,000 or 
 7,000 persons a year. Nevertheless, this tells us a lot about the 
 
 characteristics of the population that are neither reported in 
 mortality statistics or in the household interview survey. Obvi- 
 ously, we can obtain data in a physical examination that is not 
 available from a survey interview. Take the case of diabetes. How 
 large a problem is diabetes in our population? How can you 
 obtain that piece of information from the existing systems? Do 
 you go through a hospital system? That way you get only the 
 very sick; you don't get the people who never come into the 
 hospital. If you go through a death reporting system you get only 
 those who were sick enough to die. There's just no way one can 
 find out about the load this problem places on the population 
 except to go out and conduct a set of physical examinations on 
 people. 
 
 Student: What level of confidentiality do you employ in your 
 dealings with these individuals? 
 
 Dr. Perrin: Let me explain the concept of confidentiality. As a 
 statistical data-gathering agency we do not care about the 
 information concerning any single individual. In fact, we're sworn 
 to secrecy regarding the information on each individual. We 
 cannot release it in individually identifiable form. 
 
 Let me give you an example. Do you remember the 
 publicity about 6 months ago concerning the minority females 
 being sterilized in some southern States? Well, my organization 
 happens to collect the information on family planning clinics for 
 that part of the country. The Secretary of our Department asked 
 who these people were. I told him that there was no way he could 
 get any information from me about any individual in our data 
 system. This sort of confidentiality is extremely important. 
 
 We can do nothing with the information without a written 
 agreement from the individual from whom we gather that 
 information. And usually we get that written agreement. In the 
 rare case, a person will say they don't have a doctor and they 
 don't want anybody to know about their own physical condition. 
 Then we simply tell that person about his or her condition and let 
 him do what he wants. 
 
 Student: What would you do if you found someone with active 
 tuberculosis who said he didn't want the information sent to 
 anybody? 
 
 Dr. Perrin: There would be nothing we could do. 
 
 Student: And he could give his disease to other people. 
 
 Dr. Perrin: That's right. What do you think we should do? 
 
 Student: But once you see those people you sort of have a 
 moral obligation. 
 
 Dr. Perrin: Who's the moral obligation to? Is the obligation to 
 the individual or to society? When these two obligations conflict, 
 how do you resolve them? 
 
 The individuals themselves are generally quite interested 
 in the information. The question is what they do about 
 the information. The issue here is, when someone has a condi- 
 tion which is threatening society, such as a communicable 
 disease, what should we do with that information? Should 
 we hold that such information belongs to the individual for 
 
 31 
 
MONITORING THE NATION'S HEALTH 
 
 his use, or should we say that in some way that information 
 belongs to the State and we will use it as we see fit? 
 
 Student: It belongs to the individual. If he had a right to refuse 
 the examination I think he has a right to confidentiality on your 
 part. 
 
 Student: I don't agree. I think it is a matter for the State. If it's a 
 communicable disease that's going to infect the populace as a 
 whole, it's a matter for the State. And I think there should be 
 some agency that you can report it to and then the Public Health 
 Service could follow up on it and try to convince the individual 
 to be treated. 
 
 Dr. Perrin: Where do you stop? How do you define what is a 
 threat to society and what isn't? 
 
 Student: Do you have curative services in your examination 
 units? 
 
 Dr. Perrin: No, we don't. 
 
 Student: There is the Public Health Service. And there are parts 
 of the Public Health Service that give shots and so on, and so 
 these people could be referred to such an organization and they 
 would carry the ball. 
 
 Student: But, by statute, you're not a policing agency. 
 
 Dr. Perrin: Quite correct. The information we collect belongs 
 to the individual. And with his permission we will release it 
 to a physician. But without permission, it's our feeling that the 
 right to information is so fundamental to the individual that we 
 cannot assume the role of a regulatory agency and give this 
 information to anyone else. 
 
 Student: In certain ways I can see the possibility of getting 
 around that without disclosing individual names. For instance, 
 when your statistics come out you're going to show a high or low 
 incidence in certain areas, and this information could be used by 
 Public Health Services. 
 
 Dr. Perrin: Yes, that's true. That's a perfectly valid use of our 
 data. In fact, we routinely supply aggregate information on many, 
 many health factors. 
 
 Student: Can your information be subpoenaed? 
 
 Dr. Perrin: No, it cannot be, and that's important. There are very 
 few Federal agencies where that's true. You see, that's the 
 advantage of dealing with a data organization like the Census 
 Bureau or the National Center for Health Statistics. The Internal 
 Revenue Service currently can't get its hands on any of our 
 confidential data. Nobody can. In the Federal Government there 
 are a pair of opposing forces. One of these is the protection of the 
 right of the individual, and the second is the right-to-know law 
 which states that every individual has the right to know what 
 information about him is contained in a system. But more than 
 that, the public, in general, has a right to know what data are 
 involved in any of the Federal systems. 
 
 For instance, many reports of other agencies which have 
 been held confidential have been subpoenaed, and the courts have 
 
 ruled that these data must be released. These reports sometimes 
 contain information about individuals that can be potentially 
 embarrassing or harmful. Unless you have a law such as we and 
 the Census Bureau have, the right-to-know law lays you open to 
 exposure of the information in the system. 
 
 This is what the law says: "In the case of information 
 obtained in the course of health statistical activities under Section 
 304 or 306 of the Public Health Service Act, information may 
 not be published or released in other forms if the particular 
 establishment or person supplying the information or described in 
 it is identifiable, unless such establishment or person has 
 consented to its publication or release in other forms." I must 
 operate under that law, and, it seems to me, we must protect 
 the right of the individual to do with that information what he 
 sees fit. It's a very difficult decision to make. 
 
 Student: Would you supply aggregate-type information to 
 private industry for commercial use? 
 
 Dr. Perrin: Yes. We do that all the time; so does the Census 
 Bureau. The clothing industry is one of our biggest customers. 
 They like to know about people's foot size, people's body size. 
 We do a lot of anthropometric measurements and the size of the 
 U.S. citizen has been changing over time and we follow it quite 
 carefully. 
 
 You see, once the data that we collect-and this is an 
 important "other" side of the coin which I haven't empha- 
 sized—is sanitized in the sense that our confidentiality rules are 
 followed, it is available to everyone. Another very important 
 primary tenet of our data-collection agency is that the data 
 collected are made available to everyone who needs them or who 
 even thinks he needs them and the release is not controlled. 
 That's fundamental to our operation. Any data that we collect 
 are available to anyone. To control who can see and use the 
 sanitized data would be quite inconsistent with our mission. 
 
 Student: All laws are subject to test, and obviously, this one is 
 sooner or later going to be tested for the interpretation and intent 
 of the legislators. 
 
 Dr. Perrin: This is not a new section. This section actually 
 is the same one which has governed all of our work for the last 
 10 years. The Census rules were tested in court, as to whether 
 data are subject to subpoena. 
 
 Mr. Barabba: Yes, and our data's immunity to judicial sub- 
 poena was reaffirmed by Congress. 
 
 Student: In your position, do you serve at the pleasure of the 
 Secretary or are you protected by tenure? 
 
 Dr. Perrin: This was part of the controversy surrounding this 
 particular law. It's difficult to imagine that there can be so much 
 controversy around something called the National Center for 
 Health Statistics, but we generate a lot of heat, though not much 
 light on this particular issue. The Senate believes nowadays that 
 people in my position should be appointed by the President and 
 subject to confirmation hearings in the Senate. That is the case 
 with my colleague, Mr. Barabba. 
 
 Student: So, as I understand it, you do serve basically at the 
 
 32 
 
MONITORING THE NATION'S HEALTH 
 
 pleasure of the Secretary of HEW, but can he fire you if you 
 refuse to supply him with specific data he requests? 
 
 Dr. Perrin: I serve at his pleasure and he can fire me anytime he 
 wants to, but he can't get the data. 
 
 Student: I can understand, to a certain extent, why your Center 
 is in HEW; the tools are there, the medical tools, but is there any 
 particular reason for keeping it with HEW as opposed to 
 combining it with the Census Bureau and having a big statistical 
 bureau? 
 
 Dr. Perrin: Do you know anything about the statistical organiza- 
 tions in other countries of the world? Canada has something 
 called Statistics Canada which includes all of the statistical 
 activities of the Census Bureau, Labor Department, HEW, and 
 others. Statistics Canada is a centralized statistical operation for 
 its government. In fact, European governments have tended to 
 centralize their statistical activities. 
 
 We, in the United States, are somewhat unusual in that we 
 have five major independent statistical organizations recognized 
 by the Office of Management and Budget, and coordinated by the 
 Division of Statistical Policy in OMB. These are the Bureau 
 of Labor Statistics in the Department of Labor, the Agricultural 
 Reporting System in the Department of Agriculture, the Social 
 and Economic Statistics Administration in Commerce of which 
 the Census Bureau is the major component, the National Center 
 for Educational Statistics, and the National Center for Health 
 Statistics. 
 
 Recently, the Department of Justice has developed a very 
 important new data-collection activity, but each department has 
 its own in-house system coordinated with the Office of Manage- 
 ment and Budget. 
 
 Student: I'm sure you'd like to use the decennial census for a lot 
 of your health statistics, and I'm sure that Mr. Barabba is 
 probably saying that he doesn't want to make the form 17,000 
 pages long. 
 
 Dr. Perrin: More fundamental than that, the Census Bureau has 
 the only real sampling frame in the country. That is, they have 
 the most complete list of every person and household in the 
 country. We would like to be able to use that sampling frame for 
 drawing our own samples, but they won't release that informa- 
 tion to us because of their confidentiality rule. 
 
 As I have said, the Bureau carries out a large amount of our 
 sampling and survey activity. However, we occasionally run into 
 problems. Because we are a health-oriented agency we tend to ask 
 questions, or want to ask questions, in our surveys about which 
 the Census Bureau is quite shy. Recently we sent the Bureau a set 
 of questions we wanted them to ask in our household interview 
 survey, and the Bureau eliminated all the questions that had 
 anything to do with menstruation, bowels, and urine. The Census 
 Bureau did not want to endanger the response rate that it ob- 
 tains in its own surveys by being identified as an organization that 
 asks personal hygiene questions of people. 
 
 Even more important than that, 2 years ago we mounted 
 a major survey called the Survey of Family Growth, in which we 
 selected a random sample of ever-married females and explored 
 the histories of their reproduction and contraceptive use. The 
 reason was because population projections should be based as 
 much on the expected family size, the actual fertility rate, the 
 
 actual contraceptive practice, and so on, of the population. So we 
 developed a questionnaire and asked the Census Bureau to apply 
 this questionnaire for us to a random sample of the population. 
 They refused to do this for fear of jeopardizing their own data- 
 collection mechanisms by asking questions about contraceptive 
 use, and so forth. 
 
 Student: When they conduct surveys for you, do they have to 
 identify themselves only as Census Bureau people, or can they 
 say, "We're Census Bureau people and this survey is for National 
 Health Statistics"? 
 
 Mr. Barabba: Yes. If we're drawing the sample off the decennial 
 census, we'll have to identify it under Title 13 of the U.S. Code. 
 Then, we'll say this information is collected under the Center's 
 authority. 
 
 Dr. Perrin: When our interviewers go into households they say 
 that they are doing a survey for the U.S. Public Health Service 
 and they have a written document they present to the individual 
 which guarantees their confidentiality. It tells how the data will 
 be used and it's signed by me. 
 
 Student: Are there any other ways you deal with sensitive or 
 extremely personal questions? 
 
 Dr. Perrin: We are concerned about sensitivity partly because 
 response rates on sensitive issues can drop or answers can be 
 untruthful. One question we ask, for example, is, "Have you 
 ever had an abortion?" The question of whether people will 
 respond to that honestly is a very fundamental one. What we've 
 done is to introduce a new statistical technique which sounds 
 a little implausible but I'll tell you about it. It's called the 
 Randomized Response Technique, which was developed in the 
 last 4 or 5 years, whereby the respondent receives one of two 
 questions at random. As an example, the two questions might 
 be, "Were you born in the month of April?" "Have you ever 
 had an abortion?" Only the respondent knows the question 
 to which he is responding. We know the answer, but we don't 
 know which question it is for. It's rather simple to understand, 
 I think. In other words, we get a yes or no from the respondent 
 and we know the probability that it's question 1 and we know 
 the probability that it's question 2. Furthermore, we also 
 know the probability that that person was born in the month 
 of April. From these facts, we can estimate, for the popula- 
 tion we interview, the probability of abortion. But we don't 
 know for any individual in that population whether or not 
 they had an abortion. In other words, the Randomized Re- 
 sponse Technique allows us to gather information from the 
 population without having any specific identifiable information 
 from individuals about the incidence of an issue which we 
 consider to be sensitive. We employ this technique, not only 
 because we believe the question to be sensitive, but also because 
 in many cases, if we were not protected by confidentiality laws, 
 as of course most private concerns are not when they gather 
 information, we might be subject to subpoena. But, since we do 
 not have individual information, one couldn't subpoena anything 
 that was individually identifiable. 
 
 For example, let's suppose in Mr. Barabba's previous 
 incarnation as a private survey researcher, his outfit had gone out 
 and collected information on illegal activity, say, drug abuse. And 
 a person had responded, "Yes, I take drugs." Well, there would be 
 
 33 
 
MONITORING THE NATION'S HEALTH 
 
 a chance that a law enforcement agent might be able to subpoena 
 those records from a private organization. But as long as it's a 
 randomized response, there's nothing there to subpoena. So, 
 there are other reasons for the technique. 
 
 Student: We are involved in an attitude survey in my department 
 and we ran into the problem of confidentiality. It's an employee 
 attitude survey where we're asking them questions that might be 
 a little sensitive, such as, "What do you think of the department? 
 Rate it 1 to 9, very lousy to very good." The employee may 
 naturally feel that he could hurt his career if he gives the wrong 
 kind of answer. In order to solve the problem, we're using 
 another agency to collect the data. We can be sure that we're 
 getting data from every respondent, but we don't know which 
 respondent is answering which questions. It's a double coding 
 system. 
 
 Dr. Perrin: Techniques like that are important in data collection. 
 Many people who have never been involved in data collection 
 think that you can simply pass regulations and require reporting. 
 You can get answers if you require reporting, but the truthfulness 
 of the answers is often open to very serious question. 
 
 There are other data-collection mechanisms. I guess I never 
 really answered that question completely. We have a hospital 
 utilization survey where we take a random sample of hospital 
 discharges to find out what people were hospitalized for, how 
 long they stayed in the hospital, and things like that. We also have 
 a national ambulatory medical care survey, a very exciting 
 concept. For the first time we have a random sample of 
 physicians practicing in the communities of the country where we 
 go into their offices and record information on their practices. Do 
 you realize that there's no regular data-collection mechanism in 
 this country for the practice of medicine in physicians' offices 7 
 
 Student: Are you running into any resistance in the physician 
 survey? Does the AMA perceive it as an attempt by the 
 Government to establish baseline data for regulatory purposes? 
 
 Dr. Perrin: Believe me, if we did not have the support of the 
 AMA, we would not be in the field. It took us 2 years to develop 
 that support and we did it very carefully. 
 
 Student: What is the magnitude of confidentiality between the 
 practitioner and yourselves? 
 
 Dr. Perrin: We have the information on who he is and what he 
 does. Again, because of our law it is absolutely confidential and is 
 released to no one. 
 
 identify and to characterize are the types of practice by the 
 number of people seen each day, how long doctors spend with 
 each patient, how many times certain drugs are prescribed, what 
 the major complaint is. We don't even know what the distribution 
 of patient complaints is for a physician's office. We're beginning 
 togetthat information; in fact, we're also coding what the patient 
 says is wrong when he comes in. That's a terribly important thing. 
 It's not just what the doctor thinks is wrong, but what the patient 
 thinks is wrong with him that can be very important for matters 
 of health education. So, this is very exciting information which is 
 just now being processed and we hope will soon be available. 
 
 Student: When you obtain data from hospitals, do you use 
 statistical data that they, themselves, develop? 
 
 Dr. Perrin: We use the information provided by the medical 
 records group in the hospital, through the use of a standard 
 reporting form. 
 
 Student: Apparently, you have a very wide range of respondents, 
 from the administration of the facilities to the physicians, pa- 
 tients, and the nurses on duty. 
 
 Dr. Perrin: Absolutely. Each one has a special set of problems 
 that have to be worked with very carefully. That's why it 
 sometimes takes a couple of years to get these things going. 
 
 Student: In terms of public health statistics collection, is the 
 United States ahead of the game or behind, in comparison to 
 other industrialized countries? 
 
 Dr. Perrin: In some ways we're ahead. For instance, our National 
 Health Survey is a model which the rest of the world is now 
 copying. Canada is going to start one soon, Britain as well. But, 
 we're behind in another sense. Our health industry is so 
 disorganized that, in fact, the health data that are generated are 
 here, there, and everywhere. Consider, for example, Britain's 
 much more centralized health delivery system. Data flow nat- 
 urally through that system and they really have a better record 
 of the health care that's being delivered because it is paid for by 
 one source. The British, in a sense, have a better grasp on the 
 health of their people, or at least on the health care utilization of 
 their people because of their system. Now, the system we have is 
 private enterprise, really, and there's no way of pulling it all 
 together. So, in some ways we're better off and in other ways 
 we're not so well off. And the ways we're not so well off are 
 because of the lack of organization of our health care delivery 
 system, which is by design. 
 
 Student: Does the AMA either audit or review rights to the 
 information before you publish it? 
 
 Student: Your aim is to make some sense out of both the public 
 and the private sector. 
 
 Dr. Perrin: For this particular survey the answer is no, although 
 that is certainly not without precedent; we have had that 
 discussion before on other issues. We're careful not to ask 
 questions that we think are going to upset the physician. We 
 don't ask him how much money he makes. We don't ask him 
 what he charges for a particular visit because we feel the response 
 rate would drop off dramatically. Again, it's voluntary; he can 
 throw us out of his office, which some have. I get irate letters 
 occasionally. But thus far it's gone very well though very slowly 
 because of all these problems. What we hope to be able to 
 
 34 
 
 Dr. Perrin: That's our aim. It's a never-ending task and victories 
 are very small in this area. 
 
 Student: Is your aim really that or just to compile the data and 
 analyze what the situation is so the decisionmakers can attempt 
 some policy decisions. 
 
 Dr. Perrin: The difference between what you two said was that 
 the second speaker asked if we're just collecting the data and 
 making them available, whereas the first question implied that we 
 
MONITORING THE NATION'S HEALTH 
 
 are making policy judgments on the basis of those data and 
 thence recommendations? I want to make the distinction clear. 
 The answer is that we do the former. We do not become involved 
 in setting policy as a result of the information we gather. We 
 gather the information and make it available to the decision- 
 makers and the users. We haven't even fully accomplished that 
 yet. That's really why we're here, not to talk about how to 
 collect data, but how to use data. That's really the fundamental 
 question. If we are collecting data but not using them, we might 
 as well all quit right now. 
 
 Student: Who decides which data to collect, the user or you? 
 
 Dr. Perrin: Ultimately the Congress does. But the agency itself 
 sets priorities on what it wants to collect, which I approve. Then I 
 take these priorities to the Assistant Secretary for Health, to the 
 Secretary for HEW, to the Office of Management and Budget, and, 
 ultimately, to the Congress. The whole process is controlled 
 through the resource allocation mechanism. 
 
 Student: Do they decide what they want to look at by surveying 
 medical people 7 
 
 Dr. Perrin: We have a very large panel of advisors, some 400 
 people, who we contact several times a year for new ideas. Then 
 we have something called the U.S. National Committee on Vital 
 Health Statistics which is a 15-member committee advising us 
 on data collection. And we now have something called the 
 Cooperative Health Statistics System which we hope will bring us 
 in closer touch with the cities, counties, and States— the local 
 users. 
 
 It is an important and fundamental point that, as a 
 statistical agency, we should not become involved in setting 
 policy for distribution of health resources and so forth, and that 
 our task should be that of providing needed data for those who 
 make the decisions. 
 
 Student: Although you don't set policy, the reason for your 
 existence, is to collect information which will then be used to 
 make policy. To take that one step further, probably the supreme 
 policy goal in this sense is to better organize our health system, 
 which is a combination of government and private agencies. 
 
 Dr. Perrin: Absolutely. It is extremely difficult to draw the line 
 between what's data collection and policy setting. Obviously, one 
 can present data in such a way as to influence the course of 
 events. So, it's not all that easy to remain unbiased. 
 
 Student: Your system has, I think, a clear mission. Do you have 
 a performance measure for it? 
 
 Dr. Perrin: The other question is, how well are we doing our job? 
 This is extremely difficult to assess. If you are asking for a 
 performance measure for the health care system in general, you 
 need indicators. It's extremely difficult to know how well you're 
 doing a job in an organization like mine. You can convince 
 yourself you're doing a marvelous job if you talk to the right 
 people. I think we've seen evidence of that certainly in the last 
 2 months. But whether we are or not, I just don't know. 
 
 Student: Isn't it because there is no policymaking that it's very 
 difficult to evaluate? 
 
 Dr. Perrin: There certainly is a lack of something called a health 
 strategy in this country at the Federal level. Of that there's no 
 question. 
 
 At this point, we really haven't even gotten down to 
 the new thrust in the National Center for Health Statistics which 
 has occupied a lot of our time the last year or two. The tradition 
 had been that the National Center for Health Statistics produced 
 data primarily for use at the national level on events at the 
 national level. In other words, we were not able to generalize or 
 actually disaggregate our data down to the State and local levels 
 very effectively, and almost all the publications we put out 
 provided data at the national level. It was clear to many people 
 that the real need for information, in terms of magnitude, was 
 and still is at the State and local level. The greatest need is not 
 really at the Federal level, despite the fact that we need 
 information and many important decisions are being made there. 
 In terms of everyday decisions, they are made at the working 
 level in the county of Los Angeles, in the city of Tacoma, and so 
 forth, and we were not really effectively reaching the users at the 
 State and local levels. To meet this need the Center developed a 
 program called the Cooperative Health Statistics System. It was 
 previously called the Federal, State, Local Data System, which 
 was shortened to FSL, and subsequently to Fissle. We changed it 
 because we felt that the name stressed overly that we were 
 dealing solely with governments, while we wanted to deal with 
 more than simply governments. We want to deal with private 
 organizations, or whoever the data users may be. 
 
 We are attempting to develop a union of data producers, at 
 the local level, through State and regional to the Federal level, 
 and a feedback to the users at all levels. The Cooperative Health 
 Statistics System is our major new effort; an outreach program to 
 the users which I'll let Doug Williams talk about. Doug is the 
 Director of our Data Users Laboratory which is a very important 
 part of our plans for the future of the Cooperative System. Let us 
 now turn from the national scene to cooperation with State, local 
 governments, and other agencies. 
 
 Mr. Williams: One of you asked who determines what to collect, 
 and I think Dr. Perrin's answer was that the procedure is within 
 the bureaucracy for getting something out to the public. But, as 
 he said, our interests in the past have been in national statistics, 
 and what we were confronted with was the fact that other people 
 (like the State and local areas) needed good statistical informa- 
 tion for managing and planning their health programs. 
 
 The Cooperative Health Statistics System was an outgrowth 
 of the demand from the general public for some assistance in 
 trying to develop intelligence information for health planning, 
 management, and evaluation. It is a coordinated effort whereby 
 we energize, help, or in a technical way assist in devising a system 
 for the collection of data based on who can collect it most easily 
 and has the most immediate use for the data. 
 
 As you can imagine, when you're dealing with national 
 statistics, even in some of our inventory systems like health 
 facilities, by the time you publish reports on that data, some time 
 has passed. The data for national purposes are quite all right 
 because our priorities don't change as rapidly as they do in a 
 county or in a city. But for the city to sit back and wait for those 
 national statistics to apply them locally is foolish. Their problem 
 may well have changed by that time. 
 
 Local areas have had no particular statistical intelligence to 
 apply to making their decisions. So, we're trying to develop a 
 system whereby we enter into an agreement with State and local 
 
 35 
 
MONITORING THE NATION'S HEALTH 
 
 areas to produce statistics, not only just to produce them, but to 
 use them for their own purposes because of the .rapidity of 
 change at the local, State, and even the Federal levels. 
 
 Dr. Perrin has described our activities, and they've been 
 primarily on a sampling basis. Except for such things as health 
 manpower and facility inventories, our data just do not stretch 
 down to State levels, not to even mention local levels. In each 
 year's Health Interview Survey we have a sample of 200,000 
 individuals. That sounds like a lot of people and it is a very costly 
 operation, but it just doesn't get at statistics that are usable to, 
 say, local planners except from a comparative point of view. If 
 you had similar statistics in the local area, you could compare and 
 see how you stack up against the national average. 
 
 Our objective is to develop baseline statistics. We're trying 
 to get general information on the population as a whole not just 
 information through programmatic efforts like maternity and 
 child health. These are good statistics, but they only reflect those 
 people who appear at the medical door for services. You don't 
 really know what the group looks like that didn't show up. 
 
 The major aims of the Cooperative Health Statistics System 
 are to improve timeliness of data, and to allow comparability of 
 data from all sources. Quite often an individual or establishment 
 has its own private baseline statistics system and it may not 
 overlap very much with other systems. So, any time you try to 
 compare program-to-program, State-to-State, or locality-to- 
 locality, you have a big mess where you just can't relate those 
 statistics to each other. 
 
 Our goal was to reduce costs in the overall system of 
 collecting and processing data. For instance, in our vital statistics 
 system, which has been a Federal, State, and local cooperative 
 effort for 75 years, at each level they would reduce the data and 
 process it for their own needs and at the same time send in via the 
 State to the Federal people microfilmed copies of the birth or 
 death certificates themselves. Whereupon, we would code and 
 process them for our own purposes. As you can see, this is a 
 major factor in the cost of a system because we have redundant 
 data reproduction all the way up the line. What we're trying to do, 
 and I'm using this as an example because it's a very good one, is 
 to reduce the data reproduction cost by having it processed once 
 and only once, and then funnel it up through various channels 
 for a variety of uses. That's just one element of the Cooperative 
 Health Statistics System. But, as I said, it's a good example. 
 
 Therefore, we're talking about timeliness, comparability 
 and reduction in cost, and there are other factors that we can get 
 into with the development of such a cooperative effort. We do 
 have a legislative mandate for such a cooperative effort.' This is 
 part of the Public Health Service Act. It's a subsection author- 
 izing us to enter into a cooperative statistics system with States 
 and local governments or private industry. It does give us some 
 mandate for sharing in the cost of developing such a system. 
 
 At the present time, the Cooperative Health Statistics 
 System is comprised of seven components-vital statistics, man- 
 power, health facility, hospital care, household interview, ambula- 
 tory care, and long-term care statistics. The first one we've talked 
 about is vital statistics. The second is manpower statistics. 
 Throughout the country there are various agencies involved in 
 licensing health manpower, and we hope to tap into such a 
 programmatic effort and develop national statistics about where 
 and how many physicians, orthopedists, veterinarians, pharma- 
 
 1 Public Health Service Act, as amended by Public Law 91-515, sec. 210. 
 
 cists, and so forth, there are. There are 13 occupations that are 
 licensed in every State. We selected those as our manpower 
 objectives for the time being. Manpower statistics also provide us 
 a frame for getting at attributes of the practitioner and also a 
 sampling frame for further callback surveys. 
 
 Again, in all of these systems we have to coordinate with 
 the State as to their confidentiality laws, make sure that our 
 interests as far as confidentiality are protected, and, in fact, 
 monitor what exactly is being asked for, just as the Census 
 Bureau does with us. 
 
 In the health facilities area we have also an inventory of 
 facilities based primarily upon licensing. We have selected cer- 
 tain types of facilities like hospitals, nursing homes, and so 
 forth, to be included in our Health Facilities Inventory. At the 
 present time our focus really is on the three I have mentioned and 
 a fourth. The fourth area is a little harder, but we've worked that 
 out to some degree, and we're working on getting hospital care 
 statistics at the present time. Incidentally, we're very much 
 involved with the various other agencies in the Government who 
 have some interest in the collection of hospital care data. 
 
 Student: How do you define hospital care? Is it things like 
 in-patient/out-patient ratios? 
 
 Mr. Williams: In this case hospital care is the attribute of the 
 in-house population, certain diagnoses, and so forth. As Dr. Perrin 
 mentioned earlier, these records come from the medical record 
 librarian, and they list attributes of the patients. Also, staffing of 
 the hospitals and various other characteristics of the institution 
 are shown. 
 
 Dr. Perrin: But those are only in-patients. We have difficulty 
 getting our hands on information concerning out-patient care. It's 
 a much less organized concept than the hospital situation and 
 in-patient care. What we're talking about in this case is in-patient 
 care. 
 
 Student: Is the agency in close contact with the AM A in terms of 
 hospital accreditation and what they need to keep on their 
 records, and are they in contact with you in terms of what the 
 hospital provides as to specific statistics? 
 
 Mr. Williams: We are in contact more appropriately with the 
 American Hospital Association and organizations such as this. We 
 have developed data sets based upon recommendations by a panel 
 of experts, who are from, among others, the Social Security 
 Administration, private industry, and private abstracting firms. 
 Also, Professional Standards Review Organization (PSRO) has 
 recently come on the scene, and that's a very big influence on 
 hospital care statistics these days. 
 
 Dr. Perrin: You should understand that these organizations often 
 collect data for us. In fact, we contract with the American 
 Hospital Association to collect information on hospitals. They're 
 very willing to cooperate because it gives them public informa- 
 tion, plus it's a way of cutting down on multiple reporting. In 
 other words, if they deal with one Federal agency, they can get 
 all of this report into the Federal system at once. The idea is that 
 you won't have four or five different Federal agencies coming at 
 you for data. The cooperation has been very good with the 
 Hospital Association, the Medical Association, the Nursing 
 Association, and others. 
 
 36 
 
MONITORING THE NATION'S HEALTH 
 
 Mr. Williams: And that is primarily one of the complaints from 
 the hospital administrators. They feel that they're inundated with 
 all sorts of reports to fill out. Some think they have to devote a 
 full-time staff to filling out certain reports. Of course, their 
 approach is from a point of view of monitoring the system for 
 their own programmatic purposes. Our concern is getting at 
 statistical information about the attributes of either the fa- 
 cility or the patient, so other programs that have to deal with 
 the availability of services can have some information about 
 what is the current census of patients in the country and how 
 many discharges there are a year, and things like that, to make 
 certain program decisions. 
 
 In the case ot household interview statistics, we've been 
 conducting this survey for quite some time. It's very complex and 
 has worked out to be quite a science in collecting good 
 information via the interview technique. It's one of our hardest 
 ones to put into this cooperative mode, because, depending upon 
 how you ask a question and who asks it and so forth, you get a 
 variance in terms of response. Therefore, we're having a great deal 
 of what you may call shakedown cruises in various States to 
 determine how well this can be done via a cooperative effort. We 
 are doing the same thing in the other areas where we need 
 reliable, timely statistics. They're in an experimental state right 
 now. Ambulatory care statistics, our own national survey, today 
 represents only about 1,500 physicians, and we have to go to 
 great lengths to get something like an 80-percent response rate. 
 You can imagine that we really want to be up on our toes when 
 we launch anything like this in a cooperative mode. 
 
 The same thing is true for long-term care statistics. There's 
 a real hard nut to crack! How do you get information on 
 long-term care facilities? In the first place, how do you define 
 them? Then, how do you collect information on them? Anyway, 
 these are the components that now constitute the Cooperative 
 Health Statistics System. 
 
 A question was asked earlier about how we determine what 
 data to collect. Let's take the development of the health 
 manpower minimum basic data set. We convened a panel of 
 experts, to come up with initial recommendations as far as health 
 manpower was concerned. Whereupon we took the product that 
 they gave us and circulated that among two or three thousand 
 individuals at the State and local levels, universities, and the 
 Federal Government. From the responses, which were something 
 quite short of 2,000, we took the substantive ones. We 
 synthesized what they had to say, and came up with what we call 
 our minimum basic data set. It's always going to be under review, 
 but what we're trying to do is to get the State, the local, and the 
 Federal requirements, and intersect their interests to determine 
 the minimum data set that all of us have some need for. A State 
 can tack on its requirements, and I don't think the Federal 
 Government's ever going to need more than, say, the core data 
 set. This is what we're now trying to install throughout the 
 country. 
 
 Student: What is the size of the current core data set? 
 
 Mr. Williams: We've got various components and there is a 
 minimum data set for each component. There are some 30 
 questions on the health manpower minimum data set. There are 
 some 20 or so on the health facilities data set. They mostly deal 
 with the attributes of the individual or hospital. 
 
 Dr. Perrin: Our concept is that if you collect a small amount of 
 data well, apply uniform criteria, you're going to be much better 
 off, and it's going to cost you less in the long run than if you just 
 accumulate data willy nilly from various places. That's what we're 
 trying to get at, to accumulate a defined set which has 
 applicability at various levels. 
 
 Mr. Williams: Our scheme is one of trying to share needs, 
 techniques, and information about health data. We have an 
 exchange of information with the States and local areas, we 
 support each other in our aims, offer technical assistance, and 
 that assistance doesn't just go one way. We have certain research 
 efforts in the States that we benefit from. Understand that our 
 system is not so much confined to the data itself, but to 
 developing a system for retrieving information. 
 
 One of the primary vehicles for improving services is 
 through government agencies. As a matter of fact, it turns out 
 that some of our prime confederates in this undertaking are 
 government agencies working at State and local levels. Under- 
 stand that we're restricted when we talk about funding. We've got 
 some 3,000 counties in this country and when we start dealing 
 directly with the county, then we're fragmenting ourselves. 
 Therefore, what we have to do at the present time is to deal 
 primarily at the State level now and encourage the State to deal 
 at the local level. 
 
 Student: What if the State either has no interest or no resources 
 for dealing with the particular level of statistical activity that 
 both the local area and the Federal Government are interested in? 
 
 Mr. Williams: Then we have to base our approach on the 
 population we're attempting to cover. If a State has no interest, 
 we would probably try to generate some, and not go beyond the 
 State to a county level except in the area of research and 
 development. The county is a very good area for certain 
 development activities and serves as a very good lobbyist with the 
 State to generate some activity in the areas where we have a 
 common interest. Right now, what we say and what we're able to 
 do is governed by the amount of funds and capabilities that we 
 have. 
 
 Student: I can see you eventually being inundated with requests 
 for information at the local level. 
 
 Mr. Williams: Requests for information are what we're trying to 
 preclude here. We're trying to have that data generated by those 
 local people who have a need for it and then they're not going to 
 be requesting from the Center of Health Statistics data pertinent 
 to their local interest, except from the point of view that they 
 may want some comparative information to see how the rest of 
 the country stacks up. Instead of going to each county and each 
 State and getting that independently, we'll have it at a central 
 location for them and we'll process it. 
 
 These data, when processed one time, say at the local level, 
 and then sold or given to other areas, become much more 
 immediate. If we put it in machine-readable form at the county 
 level and then share that with the State, and then the State shares 
 that with the Federal Government, we've got a much more timely 
 system. 
 
 37 
 
MONITORING THE NATION'S HEALTH 
 
 Dr. Perrin: You may be interested to know that we receive, in 
 my office, approximately 2,000 requests a month for information 
 from private citizens. That runs about 25,000 requests a year, 
 many of which are not tabulated because they go directly to 
 certain divisions. So there's a large demand, and this is just the tip 
 of the iceberg. 
 
 Mr. Williams: Furthermore, the requests are for more than just 
 information. We have a tremendous demand for technical 
 assistance at the Center for Health Statistics. From the technical 
 assistance point of view, we are limited in how we can operate at 
 the present time. With budget increases and so forth, perhaps, we 
 can expand these activities and deal more closely at the local 
 level, beyond just research and development or planning and 
 development activities. In a cooperative health statistics system 
 we're not just talking about developing minimum data sets and 
 then fishing these out. We're talking about developing an 
 exchange system of cost and technical know-how. What this says 
 is that we're trying to establish health sharing systems such that 
 we can underwrite some of the costs of producing data for Fed- 
 eral interests, while at the same time the States underwrite the 
 costs for their interests. We exchange the data; we have standards, 
 definitions and procedures that we all try to adhere to for com- 
 parability. 
 
 Student: How well is the data methodology that you have within 
 a particular field of health care integrated with the broad aspects 
 of Census Bureau programs? 
 
 Dr. Perrin: Once we promulgate a certain data set for our data 
 systems, we refer it to the department level, and there it goes to a 
 standards group which considers it for adoption as HEW practice. 
 Ultimately, however, the Secretary of our Department has the 
 responsibility for setting standards in data collection. At each 
 point a decision is made as to how wide the requirements should 
 be. 
 
 Student: Do you have good cooperation from the local and State 
 agencies? 
 
 Dr. Perrin: We have good cooperation as long as we have a carrot, 
 or a stick, depending on the situation. If we provide funding, 
 most State agencies are quite agreeable to go along with the 
 standards and, of course, we do attempt to promise cutting down 
 Federal requirements in data. One of the biggest things that we 
 can do is to say, "Look, if we agree on how to do it once and for 
 all, and do it right, you won't have all these programs coming at 
 you for different reporting requirements." They're very receptive 
 to that. 
 
 Student: Are there any States with poor health care that do not 
 wish to cooperate with you? 
 
 Dr. Perrin: I don't know of any instances in which that's the 
 case. However, in most States where health care is somewhat 
 below the national average, the ability to gather data on it is also 
 below the national average. The two problems seem to be highly 
 correlated. We generally lack information on areas where health 
 care is the poorest. 
 
 Mr. Williams: In that particular area, statistics for damning 
 purposes isn't our game. I don't think the State would be 
 
 38 
 
 reluctant to provide us with the information for our purposes 
 which are really finding the general characteristics of the 
 population. However, it doesn't prevent some data user from 
 taking our findings and making a case for or against a State. Our 
 system doesn't directly have that intent. However, a lot of people 
 think that anything that goes to the Federal Government is all 
 coordinated with the Census and IRS and with this, that, and the 
 other thing. So, we do run up against that sort of attitude 
 occasionally. We don't have any experience directly where people 
 want to protect themselves so they don't furnish data. 
 
 Student: Perhaps it's that you haven't discovered it yet. 
 
 Mr. Williams: That very well may be. Someone asked earlier what 
 participation we have with the States. More, really, than we can 
 handle right now because of our budget. We've had a tremendous 
 response from the one Request for Proposals that we've put out 
 on implementing these four data systems: vital statistics, man- 
 power, facilities, and hospital care. 
 
 Student: Your last point was cost sharing with the States. I can 
 see a lot of repercussions there. You can say that if they help out, 
 the Federal Government can have a real fine health data system. 
 But suppose they feel they are paying for something they don't 
 need. 
 
 Mr. Williams: What we're trying to do is, in a sense, purchase 
 information from the States. We want to underwrite our fair 
 share of the production of this data. 
 
 Student: Is this prorated? 
 
 Mr. Williams: Now it's really flying by the seat of our pants 
 because we don't have any experience as to what some of these 
 data systems cost. What we're doing is soliciting contracts 
 through a competitive contract mechanism. In the future, we 
 hope that through one of our advisory committees to the 
 Cooperative Health Statistics System we can come up with a 
 funding formula. 
 
 Here are some of the activities of the division of the 
 Cooperative Health Statistics System. We have research and 
 development activities that have been ongoing for 3 years and 
 we plan to increase them in the future. We develop the core data 
 sets and try to promote them and get interested parties to collect 
 them, and to suggest changes. This will continue to go on for 
 some period of time and probably we'll never really set it in 
 concrete. But we hope there will be minimum change for 
 temporal references and for analysis. We have a technical 
 assistance program where we actually assist the States in 
 implementing a system such as the one we've been talking about. 
 We have training and personnel development ability and capabi- 
 lity. We have the applied Statistics Training Institute, which gives 
 short courses. In the Data Use Lab we hope to launch a 
 longer term training session. We also invite, where space and 
 personnel are available, interested parties to come into the center 
 and nose around to find out what we're doing and to take 
 advantage of some of our experience, and we're also doing that 
 very same thing on the outside. We're going out into the States, 
 looking at some of the activities in the States and trying to adapt 
 what is better than our procedures, especially in the area of 
 analysis and utilization of data. 
 
 We have a communications activity, the product of which 
 
MONITORING THE NATION'S HEALTH 
 
 has been primarily newsletters. These generate quite a bit of 
 response and interest in the Cooperative Health Statistics pro- 
 gram. You might well imagine that we have generated quite an 
 extensive mailing list over the past 13 years. On the list are vital 
 statistic registrars, public health schools, epidemiologists, univer- 
 sity vital statistics programs, health officers, associations, etc. 
 
 Dr. Perrin: But, I think we have to be honest and say that data 
 dissemination is one of our greatest problems. Quite frequently, 
 we don't know where to disseminate things. Often we have some 
 very good material that doesn't get to the right people. 
 
 Student: Do you send your data and information regarding 
 health facilities and hospital care to hospitals? 
 
 Mr. Williams: We have personnel traveling around the country 
 talking to various interest groups. We request them to write to us 
 and specify where their interest may be by generic category. 
 Then we come up with a mailing list for certain types of informa- 
 tion. We produce monographs that are analyses in a comparative 
 sense of the various types of data we produce. Through our own 
 efforts, others, such as comprehensive health planning agencies, 
 receive our literature. If we develop something that we recognize 
 they have a need for, then we'll send that out to the Federal 
 comprehensive health planning group who will, then, in turn, 
 notify all of their people. All vital statistics registrars in the 
 country get copies of our vital statistics volumes. We also send 
 them to libraries— certain libraries like those in schools of public 
 health, and programs where we can identify an interest in some of 
 the data we produce. But, by and large, distribution is on a 
 request basis. 
 
 The last thing I want to talk about is the Data Use 
 
 Laboratory and the use of the data itself. Those people interested 
 in health data are generally handicapped by the fact that there are 
 usually no analytical types on their staff. What we're trying to do 
 in the Data Use Lab is to develop a procedure and some examples 
 of how to use data for certain settings. Understand that we have 
 to protect ourselves from being identified with any programmatic 
 interests, but we can concentrate on methodology. We can 
 concentrate on the packaging of information. So, in the Data Use 
 Lab that's our focus, to try to identify innovative methods for 
 using data. 
 
 We're also trying to develop packaging techniques so that 
 we can make data more understandable to various types of people 
 who have to make the decisions, for instance, legislators. You go 
 at them with a giant ream of computer printouts, and you know 
 where it's going to go, and you know what kind of names you're 
 going to be called for having brought it in. But if you produce for 
 legislatures maps, various types of graphs, and some summary 
 analytic statistics, you do have a channel of communication. For 
 instance, if you talk about the odds of a person getting a specific 
 disease, you're talking a language that people can understand. 
 
 Therefore, what we're trying to do in the Data Use Lab is 
 come up with a few procedures whereby you can go directly from 
 raw data to understandable formats such as computer mapping. 
 Of course, we're talking about expensive equipment, so a small 
 county in Mississippi or somewhere like that still might do it by 
 hand. But they can apply the procedure to make the data more 
 understandable. 
 
 Student: How current is your information? I know that after a 
 
 holiday weekend the Highway Patrol tells us how many traffic 
 deaths there were for that weekend. How current is the 
 information that you can provide? 
 
 Mr. Williams: That's one of our big problems right now. Our data 
 is not that current. For example, "vital statistics" is a couple of 
 years after the fact. That's one of our big problems and that's one 
 of the things we're trying to answer with this Cooperative Health 
 Statistics Program. We don't have weekly reporting except in vital 
 statistics, where we also have preliminary reports monthly by 
 State 
 
 Take the Health Examination Survey; it comes out several 
 years after the cycle of examinations because of the number of 
 individuals examined, and we have to go through certain 
 probability cycles before we can produce any statistical informa- 
 tion representative of a larger population. It takes us a full year to 
 get through a probability cycle, after which our analysts can take 
 over and start producing monographs. In something like the 
 Health Examination Survey, or the Health Interview Survey, 
 you're talking about complete data a couple of years after the 
 fact. That's what we're trying to overcome with our local 
 emphasis program— to develop the data at the level where it's 
 most needed and have it immediately available for us. 
 
 I have a couple of illustrative examples of the types of 
 things that we're doing now that are of interest to local areas and 
 can be applied there. One is what we call our Standardized 
 Microdata Tape Transcripts. This is elementary information with 
 all descriptive and identifying information removed, of course. 
 For instance, we often package a piece of information that may 
 have all sorts of data in it unnecessary for the user's immediate 
 needs, and in that case one might want basic data entries that 
 answer a specific interest. 
 
 The Census Bureau does a beautiful job of packaging its 
 information in microdata form. We've only begun to do that, and 
 really haven't developed the capability that we will have in the 
 future. Also, because of our data being limited primarily to 
 national interest and national programs, and because of the 
 clamor from States and localities for the type of information we 
 produce, we've come up with a fudge factor system for 
 developing information that can, in some way, give you a first- 
 line intelligence on what may be happening in your State. We do 
 this by taking the national rates and applying these to the 
 demographic structure of a State. We call it synthetic estimates 
 because this is a manipulation of our real estimates that have 
 known reliability. The synthetic estimates have no known 
 reliability; they're simply a quick look at what may be happening. 
 
 Dr. Perrin: Let's spend a minute or two on the topic of social 
 indicators. We haven't really said anything about health indica- 
 tors. The reason is we're still learning about them. I invite you to 
 think of the concept of health indicators and how useful they 
 would be if we were able to generate some meaningful ones. It 
 may appear on the surface to be rather easy for us as purveyors of 
 health data to provide health indicators. 
 
 Can you think of some things that would be useful as 
 health indicators that might be useful data to collect? Classically, 
 there have been certain ones. Do you know what the traditional 
 health indicator has been for years? Have you ever heard of the 
 infant mortality rate? That's a health indicator of sorts. It's a 
 measurement of the proportion of the infants that die within the 
 first few weeks of life. For many years people have felt that if 
 
 39 
 
MONITORING THE NATION'S HEALTH 
 
 this rate is high, health care is bad. If it goes down, health care 
 gets better, and so forth. What do you think about this as a health 
 indicator? 
 
 You have to ask yourself, what is it that you want to 
 indicate? What do you want your indicator to do? I'm often 
 asked for a measure which will make it possible to determine the 
 effectiveness of health care delivery systems. That's a large 
 problem. If we had an indicator which we could follow over time 
 which would fluctuate as a result of programmatic changes, we 
 might possibly be able to answer many policy questions. Can you 
 think of any other possible indicators of health? "Causes of 
 death," you say. These do change over time. We've seen a dra- 
 matic change from the acute infectious diseases in the last 30 
 years down to the more chronic diseases today. In fact, in the last 
 couple or 3 months, we have shown that deaths due to heart 
 disease have decreased over the last 5 or 6 years. 
 
 our indicators lack sensitivity, that is, they don't fluctuate as ap- 
 propriately, or lack specificity in that they fluctuate as a result of 
 things that appear to be quite independent of health care. 
 
 Student: Do indicators come out of the PSRO's because the 
 physicians are supposed to determine the quality of care? 
 
 Dr. Perrin: PSRO's will be asking the question, is the quality of 
 care sufficient in this area? They will have to define what they 
 mean by "quality of care" and they'll have to measure that. Re- 
 search into the quality of care will yield indicators on quality of 
 care, I'm certain. I don't know what those indicators will be. 
 Length of stay in hospital is one of them, of course. Because if 
 you have an area in which a person goes into the hospital and 
 stays much too long as a result of that condition, they tend to 
 ascribe that to poor medical care. 
 
 Student: Since coronary care units? 
 
 Dr. Perrin: Since coronary care units. You are making the 
 lightning jump from the fact that death due to heart disease is 
 going down to the fact that it is probably due to our having 
 coronary care units. I'm not willing to go that far with you. 
 That's an indicator. It's certainly one of the ones you'd have to 
 look at. We simply produce the data, we don't interpret it. 
 
 Frankly, we are not that well equipped with health indica- 
 tors in this country. By the time you have talked about infant 
 mortality and average length of stay in the hospital, you've just 
 about exhausted them. And, maybe these indicators are not sensi- 
 tive enough. 
 
 Student: What about group medical practices? Is that going to 
 help you at all in getting more data? 
 
 Dr. Perrin: The question is not getting the data; the question is 
 defining and conceptualizing the indicators, the variables, putting 
 the variables together into a meaningful set of indicators or a 
 meaningful index. 
 
 Student: You made the statement that perhaps the indicators are 
 not sensitive enough. In the development of indicators, isn't there 
 a danger that whoever defines those indicators might come up 
 with some that say that the U.S. ranks number one? 
 
 Dr. Perrin: Right, they have to be careful about that. Many of 
 
 Student: How far advanced is the idea of a medical audit 7 
 
 Dr. Perrin: Quite far. I think it is generally accepted by hospitals. 
 
 Student: Most of that is being done on the professional side. 
 Now, what's being done for the consumer? How are they having 
 access to, say, medical audit records, or the PSRO files? 
 
 Dr. Perrin: They do not have access to them because they are 
 considered part of our or the physician's confidential files. A hos- 
 pital is considered a committee of physicians. I think if you look 
 at it on the other side you would have law suits like you can't 
 believe. Everyone could come in and say that a given level was 
 the standard and it had been missed by so much, and maybe it 
 meant 2 more days of hospitalization and everybody reacts dif- 
 ferently to treatment. Your hospitalization would become in- 
 credibly expensive because of malpractice suits. 
 
 As soon as you start to apply standards for quality of care, 
 you force people to do things that they often would not do and 
 the cost of medical care goes up dramatically. For example, the 
 reason so many of us have to have so many lab tests is not be- 
 cause, in fact, they're really necessary, but because of the profes- 
 sion's feeling that if they don't do these, one may be attacked 
 with a malpractice suit. A lot of the price that we pay for medical 
 care is a result of the physician making certain that he meets all 
 of the possible kinds of standards that might exist. You have 
 costs pushing in one direction and quality pushing in the other, 
 and the two seem to be in conflict. 
 
 40 
 
IEASURING NATIONAL WEALTH AND WELL-BEING 
 
 Sidney L. Jones, Deputy Assistant to the President 
 
 "I wholeheartedly agree that such measures of human welfare as our health, our 
 legal system, our environment and our resources are essential, and they are of 
 fundamental importance in almost every significant governmental decision. When govern- 
 ment officials attempt to make a policy decision they frequently discover that the basic 
 information is not available. A more comprehensive set of statistics covering various social 
 questions is definitely needed." 
 
 Sidney Jones 
 
 Mr. Barabba: This afternoon we're going to hear from Dr. Sidney 
 L. Jones, who is currently serving as the Deputy to the Counselor 
 to the President for Economic Policy. Prior to this position, he 
 served as the Assistant Secretary for Economic Affairs with the 
 Department of Commerce. In that position he was my immediate 
 superior, working with the Bureau of Economic Analysis and the 
 Bureau of the Census. Before that, he served in other Government 
 posts, including Senior Economist and Special Assistant to the 
 Chairman of the Council of Economic Advisers and Minister- 
 Counselor for Economic Affairs for NATO. Prior to that he was a 
 professor at the University of Michigan and Northwestern 
 University. He is a most distinguished economist and, in addition 
 to that, he's a pretty nice guy to work for. He's not as nice as he 
 looks. Sometimes, he gets quite tough. So don't let his pleasant 
 demeanor fool you. 
 
 Dr. Jones: I think it would be like the emperor with no clothes if 
 I did not comment on today's events before discussing some of 
 my views on the role of economic statistics. I will then talk about 
 the Gross National Product accounts, and what relationship they 
 might have with the subject of social indicators. 
 
 This is my fifth year in Government service, beginning in 
 1969 with the Council of Economic Advisers. Obviously, the past 
 month has been one of the most difficult phases of that tenure. 
 Nevertheless, I think it has been a great credit to our Nation, to 
 our institutions, and to our Government that over the last 2 
 years, and particularly the past several months, so much good 
 work has been accomplished under the most difficult of 
 circumstances. Even as the drum beat of events has accelerated 
 over the last month, I have been impressed that the functions of 
 Government have proceeded reasonably well. Some of my 
 activities have not proceeded as well as they might have, but in 
 the main, I would say that the Government has done well and that 
 its achievement is a credit to all the people involved. The 
 important thing is that we have discovered that our institutions 
 really do work. 
 
 In summarizing my views on economic statistics, I will rely 
 on personal experiences that extend back to 1969 when I was at 
 the Council of Economic Advisers. Over the past year, it has been 
 my privilege to work with what are probably the two most 
 professional organizations in our Government: the Bureau of the 
 Census and the Bureau of Economic Analysis. 
 
 The Bureau of the Census is a major collector of informa- 
 tion and the Bureau of Economic Analysis has major analytical 
 responsibilities. I believe we have had a happy marriage. The 
 Bureau of Economic Analysis is a much smaller organization 
 
 (about 800 employees) as opposed to the larger unit, the Bureau 
 of the Census, which employs about 7,000 people. About half of 
 the Census employees are enumerators. But, the two bureaus do 
 work well together. The Bureau of the Census collects the 
 economic and social information, which then is used by the 
 Bureau of Economic Analysis, along with inputs from many other 
 agencies, to develop the National Economic Accounts. 
 
 Economists have been involved in efforts to develop 
 National Economic Accounts for at least several hundred years. 
 In the 1700's the British Government developed a set of accounts 
 to measure the results of the industrial revolution. In the 1920's 
 the U.S. Government included a group of economists who 
 analyzed rail car loadings and pig iron output to determine what 
 was going on in the economy. That group was the predecessor of 
 the Bureau of Economic Analysis. 
 
 The Great Depression of the 1930's highlighted the need 
 for a more sophisticated system of economic intelligence. 
 Fortunately, it was about that time that the intellectual inputs 
 were forthcoming from various sources. The 1940's accelerated 
 the trend toward an integrated set of economic statistics because 
 a major part of the wartime mobilization effort involved the 
 allocation of resources, and that is what economics is all about. 
 By the end of World War II, we had what we now would 
 recognize as the National Economic Accounts; both the Gross 
 National Product, which is largely the product side of it, and the 
 National Income Accounts, which are the results of those 
 activities. Since then a series of modifications have been made to 
 improve the accounts to their present level of refinement. We 
 clearly recognize that additional improvement is required and the 
 Bureau of Economic Analysis is more aware of that need than 
 any other group. 
 
 There are several points that have to be emphasized. First, 
 the accounts are used to measure production or output. Second, 
 they are focused on the output of goods and services which can be 
 measured in financial terms. Third, inputs, the costs of those 
 elements that result in a finished good or service, are measured 
 according to whether or not they take place in the market system. 
 These three points are quite important for understanding what we 
 are trying to do, as opposed to what various social measuring 
 techniques might attempt to achieve. Because the accounts are 
 output-oriented, they summarize the final values of goods and 
 services, not their intermediate values. To make it even more 
 complicated, the measures must fit into finite time periods of 
 either 3 months or a year. 
 
 The basic concept of the accounts is rather simple. One 
 adds up the financial value of what is consumed by individuals. 
 
 41 
 
MEASURING NATIONAL WEALTH AND WELL-BEING 
 
 outlays by the Government and the business sector for invest- 
 ment and procurement and the net effect of our international 
 transactions. The summary total of these major sectors becomes 
 the Gross National Product. Underlying that total, however, is a 
 complicated set of assumptions and many very complex defini- 
 tions. In most cases, the values used involve specific transaction 
 prices reported in the market place. However, there are some 
 situations in which estimated values must be used. For example, 
 we estimate the value that a homeowner would have to pay to 
 rent a residence and the value of the food that a farmer may 
 produce for personal consumption at home. We estimate the 
 value of commercial bank transactions because there are many 
 services in the commercial banking industry which, because the 
 interest rate carries the burden, are not directly charged. 
 Basically, the Gross National Product attempts to record the 
 transaction value of our Nation's economic activities. 
 
 For many years these accounts have provided a useful 
 record of national economic events. Even more important, the 
 National Economic Accounts are of crucial value in analyzing 
 current economic change. The creation of national policy is 
 dependent upon such analysis. In the private sector, individual 
 companies and consumers need to understand what is happening 
 in the economy. Decisions at all levels depend upon this set of 
 detailed statistics. 
 
 Gross National Weil-Being 
 
 Beyond economic measures, there are many other values 
 that can be generally referred to as welfare goals. There is 
 certainly more to life than the output of goods and services as 
 measured in the GNP accounts. Obviously, it is extremely 
 important and interesting to know as much as possible about the 
 welfare status of any society. Therefore, many analysts have 
 advocated the development of additional statistics to measure the 
 "gross national welfare." Part of this interest results from a basic 
 concern that the existing GNP accounts, as we know them, may 
 not be the best way to measure welfare values. The existing GNP 
 accounts clearly do not include a number of things which are very 
 relevant to the quality of life or the welfare of our society. I 
 would agree wholeheartedly such measures of human welfare as 
 our health, our legal system, our environment, and our resources 
 are essential, and they are of fundamental importance in almost 
 every significant governmental decision. When government offi- 
 cials attempt to make a policy decision, they frequently discover 
 that basic information is not available A more comprehensive set 
 of statistics covering various social questions is definitely needed. 
 
 My basic point is that we have a variety of valuable 
 statistical measures. Some are combined to report the gross 
 national output of goods and services. Others are used separately 
 to evaluate the status of specific welfare needs. Each set of 
 statistics is important and useful. But having said that, I think it 
 would be very unfortunate to confuse these two issues. That is, 
 we have a measure of output of goods and services, stated in 
 financial terms and tied to market transactions, which claims to 
 be nothing more than that. The economic significance of such 
 statistics should not be confused or compromised by trying 
 to convert the Gross National Product into more than it really is. 
 The GNP is not a measure of our national welfare. The fact that 
 the GNP may rise from $1,288 trillion in 1973 to $1.4 trillion in 
 
 1974 tells us nothing about the welfare needs of the country 
 the various causes of human despair and happiness, the unfortu- 
 nate animosity between races, the quality of our homes and 
 families, general health conditions, or many other welfare issues. 
 The GNP figures do give us a hint as to what employment levels 
 are likely to be, and the availability of work opportunities 
 obviously does influence every other social issue. Yet, to attempt 
 to make the GNP accounts into a more comprehensive measure of 
 social welfare than they are would probably detract from the 
 existing important advantages provided by these GNP statistics. 
 
 Critics of the current framework of GNP accounts argue 
 that they are much too narrow because they fail to measure the 
 current benefits resulting from previous capital investments in 
 buildings, machines, education, health, etc. Although all of these 
 investments provide a source of future benefits they are not 
 reported in any current Gross National Product because only the 
 current output of goods and services is included. In other words, 
 if you erect a building in 1973 the initial outlays involved are 
 reported only in the GNP for 1973, even though the building 
 provides shelter or office facilities for many subsequent years. 
 The GNP will not reflect the ongoing output of a structure over 
 the coming years. 
 
 Another major criticism of the GNP accounts concerns the 
 procedures used for reporting on economic activities involving the 
 Nation's environment. The Bureau of the Census has a major 
 project underway to identify the environmental injuries asso- 
 ciated with economic activities. Next year it hopes to collect 
 information on the capital investments for pollution abatement 
 so that some cost/benefit comparisons can be prepared. The 
 Bureau of Economic Analysis has a team of researchers working 
 on the complex definitional problems involved so the purchases 
 of equipment and payrolls committed to pollution abatement can 
 be identified. But it is difficult to classify activities which avoid 
 pollution. For example, assume that a very simple economy 
 commits all of its material and human resources to the produc- 
 tion of 1,000 loaves of bread. If that simple society subsequently 
 decides that the existing production processes have to be adjusted 
 to avoid pollution, then the number of loaves produced might 
 decline to 950 loaves. In this simplistic example, the total GNP 
 would be reduced even though society continued to achieve its 
 identified goals. 
 
 This abstract example can be further complicated by 
 considering two additional factors. First, if the economy had 
 additional human and material resources available beyond those 
 used for satisfying the existing demand for bread, it might be 
 possible to continue to produce the 1,000 loaves of bread and 
 achieve pollution abatement goals as well. A second situation 
 might develop where the society decided to use some of the 
 resources previously used to produce bread for pollution abate- 
 ment and correction efforts. Under this set of assumptions the 
 output would consist of 950 loaves of bread and the "services" 
 created by the other activities. 
 
 These examples are exceedingly simplistic and fail to 
 dramatize the difficult definitional problems involved. The 
 decisions about pollution abatement are not dependent upon 
 GNP measurement considerations but these problems do offset 
 our economic accounts. If resources are available to work on 
 pollution abatement, without reducing the existing output of 
 goods and services, then the services provided have to be properly 
 defined and measured for inclusion in the economic accounts. 
 The capital investment in equipment can already be measured 
 
 42 
 
MEASURING NATIONAL WEALTH AND WELL-BEING 
 
 under existing GNP definitions. There is a value to the current 
 analytical effort because it indicates that resources are diverted 
 when pollution abatement efforts begin. If resources are already 
 fully utilized, then current output declines and the offsetting 
 "services" have to be identified. 
 
 The basic message of economics is that resources are limited 
 and must be carefully allocated to a long list of competing 
 demands. 
 
 Measurement Difficulties 
 
 Student: Wouldn't the Government simply identify the value of 
 the pollution abatement services? 
 
 Dr. Jones: One of the great myths is that there is a central 
 government agency that can identify and solve (including the 
 financing) all of society's problems. There is a Congressional 
 appropriations system which can allocate funds to pollution 
 abatement programs, but the definitional and measurement 
 problems of preparing the National Economic Accounts still 
 remains. The value of cleaner air and water is a difficult concept 
 to measure. A similar problem involves the measurement of the 
 actual deterioration of the environment resulting from economic 
 activity. The economic benefit of a pollution abatement effort is 
 not the same thing as the input or cost value created by the 
 activity. 
 
 Student: Are you saying that the benefits could actually be more 
 than the cost of preventing the pollution? 
 
 Dr. Jones: My assumption is that the financial value of the 
 benefits would be more than the abatement expenditures, but it 
 is extremely difficult to impute a value to these benefits. The 
 other economic activities involving consumption, investment, 
 government outlays, and the net value of foreign exchanges of 
 goods and services all have an explicit transaction price. 
 
 Student: Can't you impute certain types of values? For example, 
 in the very simple analogy you just referred to, isn't there a 
 sacrifice in the economic decision that is made? Isn't it a question 
 of who's going to pay for the avoidance of pollution? You 
 mentioned an obvious example involving consumers who decide 
 they will accept fewer loaves of bread at a higher per unit cost if 
 human and material resources are already fully employed, so that 
 the diversion of such resources into preventing pollution reduces 
 the output of other goods and services such as bread. 
 
 Dr. Jones: Yes, the value of pollution abatement could be 
 crudely estimated using the methods suggested. But the problem 
 becomes more complex as we turn to other alternatives such as 
 public parks or cultural opportunities. 
 
 Student: Wouldn't it be possible to measure the capital expendi- 
 tures required to enable operating enterprises to comply with 
 whatever air pollution and water pollution regulations the 
 Government decrees? It seems to me that these capital outlays 
 and incremental operating costs could be identified as they are 
 absorbed. 
 
 Dr. Jones: The Bureau of Economic Analysis has been working 
 on this exact problem for some time and its report will be 
 forthcoming soon. The Bureau of the Census has a current project 
 that is trying to measure the value of capital investments for 
 
 pollution abatement purposes. However, the problem remains of 
 estimating the resulting benefits. In order to calculate a cost/ 
 benefit analysis both pieces of information are required. 
 
 An interesting example of the problems involved in this 
 type of analysis was discussed in chapter four of the 1970 
 Economic Report of the President, prepared by the Council of 
 Economic Advisers. The issue involved air pollution standards and 
 the effects of their enforcement upon the sulfur emissions of 
 various polluters, particularly smelters in the Northwest. There 
 are two different ways that these standards might be enforced. 
 One approach is to have the Environmental Protection Agency 
 establish an arbitrary standard requiring everybody to cut out 95 
 percent of their air polluting emissions. For most homeowners, 
 utilities, and many companies that operating adjustment would 
 be feasible without excessive costs. But, for smelters, there is no 
 way that they can do it given existing technology and costs. An 
 alternative approach to achieve the overall goal of a 95-percent 
 reduction in total air emissions would be to permit varying levels 
 of performance as long as the total correction is achieved. In the 
 example presented by economists a pollution emission tax would 
 create an economic incentive to reduce pollution. For example, a 
 so-called "pollution tax" of 3 cents per pound on sulfur emissions 
 might be imposed. For a large utility with high stacks and the 
 technology to trap the emissions, it might be economically 
 attractive to install pollution abatement equipment to avoid the 3 
 cents per pound tax. However, if the costs of preventing pollution 
 for a specific company, such as a smelter, exceed the tax to be 
 paid, the firm may decide to pay the penalty. Nevertheless, the 
 overall goal of 95-percent improvement in the air can be achieved 
 by having other utilities, manufacturers, and individuals eliminate 
 100 percent of their pollution. The tax revenues, of course, could 
 be used for research and development to create the necessary 
 technology for the smelter to reduce its pollution and for 
 restoration of the environment. 
 
 There is a difference between the regulatory approach and 
 the economic approach. Unfortunately, some critics are unwilling 
 to consider the second because they mistakenly consider it a 
 license to pollute. To the contrary, the economic approach 
 merely allows a more flexible means of achieving the same overall 
 goal. 
 
 Student: In your explanation you said that if a company can use 
 antipollution equipment at a cost of less than the penalty tax it 
 would have an incentive to do so. You defined the whole scheme 
 on a pure economic cost basis. To me, that means that if you can 
 afford it, you can pollute, and if you can't, you don't pollute. 
 
 Dr. Jones: No. What I'm saying is that if it has the proper 
 economic incentives, operating through available technology or 
 improved efficiency, to adjust at a cost less than the per pound 
 tax, then it would pay the company to do it. If not, it would pay 
 the individual firm to continue its existing program. But the key 
 point is that the total pollution abatement goals would be 
 achieved because those with the economic capabilities to reduce 
 pollution would be specifically motivated to do so. The blunt fact 
 is that technology is not yet available to permit every company to 
 eliminate 95 percent of its unwanted emissions. 
 
 Student: Why have an option? If it's worth going after the 95- 
 percent goal, why not make everyone comply rather than merely 
 fining them? 
 
 43 
 
MEASURING NATIONAL WEALTH AND WELL-BEING 
 
 Dr. Jones: You are assuming that some philosopher king can 
 decide what the right adjustment price should be and that you 
 really don't care that some smelters go out of business. 
 
 Student: We're talking about saving the environment. 
 
 Dr. Jones: No, we're not. We are talking about the total goals of 
 society which include providing jobs, expanding capital invest 
 ment, providing the goods and services that people want, assisting 
 other nations, and many other things. 
 
 Student: Aren't we talking about some sacrifices being made? 
 
 Dr. Jones: Yes, but we should be talking about doing it 
 intelligently. Remember that the pollution abatement goal is to 
 eliminate 95 percent of the unwanted emissions. By having those 
 individuals and companies that are capable of eliminating 100 
 percent of their pollution do so, the overall goal can be achieved 
 without destroying those specific firms that cannot adjust. It 
 should also be noted that the pollution tax revenues could be 
 used to restore the environment. 
 
 Student: Who decides that the goal should be a 95-percent 
 reduction? 
 
 Dr. Jones: This is an arbitrary rule promulgated by the Federal 
 Government. 
 
 Student: If society can achieve a 95-percent reduction without 
 asking the smelters to reduce pollution, it may be that the goal is 
 too low. 
 
 Dr. Jones: Yes, that is true. My explanation was only intended to 
 suggest two different methods of achieving the identified objec- 
 tive. The first approach is to issue a Federal regulation which says 
 that everyone must reduce his air emissions by 95 percent. The 
 alternative is to develop economic incentives to encourage a more 
 efficient adjustment process. 
 
 Student: Do you favor the second approach primarily because 
 it's more economically viable and efficient and because it allows 
 you to calibrate the equilibrium point more precisely in a 
 financial sense? 
 
 Dr. Jones: Exactly. The analyst may not know the proper 
 emission tax to charge, but the economic approach permits 
 adjustments to continue until the equilibrium point is deter- 
 mined. In this way economic and environmental concerns can 
 both be recognized. 
 
 Another criticism of the existing GNP accounts involves the 
 absence of some imputed value for service performed in the 
 home. The financial value of domestic services are included in the 
 National Economic Accounts if they are performed by a non- 
 family member. But in the grotesque example that economists 
 love to use, envision a situation where a man marries his 
 housekeeper. In the current National Economic Accounts, the 
 GNP at that point would decrease because the services of the 
 housekeeper would be eliminated and we have no way to impute 
 the value of the services now being performed by a wife. 
 
 Student: Why not impute a value to the services of housewives? 
 
 Dr. Jones: How would you do that? 
 
 Student: By assigning the market value of their services to the 
 time spent on homemaking duties. 
 
 Dr. Jones: But what is the market value of their time? 
 
 Student: It would be $2.00 an hour, $3.00 an hour or whatever 
 it is. 
 
 Dr. Jones: But whatever is it? What specific rate of compensation 
 would you use? 
 
 Student: You could use the rate that is paid to people who are 
 working at housekeeping duties and getting paid. You could use 
 those standards. 
 
 Dr. Jones: Are you suggesting that the household worker is 
 interchangeable with a housewife? 
 
 Student: Are there any benefits for being a wife? 
 
 Dr. Jones: My own view is that this is the most important calling 
 in our society. But that is a value judgment. The problem for us 
 to consider is whether or not the activities of a wife can be valued 
 using the financial measures applied to payment of domestic help. 
 
 Student: Aren't they doing the same functions? They are 
 performing many services. 
 
 Dr. Jones: My wife works much harder than I do. I don't deny 
 that there is a fantastic service being performed. But, I'm not 
 asking that question. I am asking: What services are performed, 
 which of those services should be measured in financial terms, 
 and what rate of imputed financial payment should be used? For 
 example, should the imputed value of every wife be equal? 
 Should the imputed value change over time? Should the imputed 
 value differ according to age, other part-time functions, number 
 of children, other activities outside of the home, etc.? In short, the 
 determination of an imputed charge seems to be literally 
 impossible. This does not mean that the services of wives are not 
 of great value to society as well as to their families and husbands. 
 
 Student: But you do pay your wife something. You provide 
 services of a protective sort— food, shelter, and all that. You have 
 an income that accrues value to both of you unless you abstain 
 from giving your services. 
 
 Dr. Jones: Are you suggesting that I should use as a proxy what a 
 domestic worker earns? 
 
 Student: Why not? 
 
 Dr. Jones: Because a wife is something much different than a 
 domestic worker. 
 
 Student: Now, are you just being facetious? 
 
 Dr. Jones: No. I'm saying that it sounds so easy to accept that 
 simple substitution concept, but the problems referred to above 
 
 44 
 
MEASURING NATIONAL WEALTH AND WELL-BEING 
 
 make it impossible to figure out how to do it. As a fall-back 
 position I suppose we could impute a value for only those manual 
 services associated with domestic service as distinguished from the 
 diverse contributions of a woman in the role of a mother and 
 wife. That would narrow the complexity of the valuation 
 problem, but it is still impossible to identify the average "output" 
 of a wife in functions limited by definition to domestic service. 
 
 Student: Hasn't it been defined, though, as $12,000 a year? 
 
 Dr. Jones: That figure was approximately the average total 
 family income in 1973. 
 
 Student: But a wife is a general practitioner. She is not a 
 specialist in terms of the type of labor performed. 
 
 Dr. Jones: I certainly agree, and that point further complicates 
 the valuation process. 
 
 Student: In economic terms, what is so mind-boggling about 
 estimating her total services if you change the situation to assume 
 that your wife, or anybody's wife, were to perform the same 
 household services of laundering, cooking, or house cleaning, in 
 somebody else's home for a price? 
 
 Dr. Jones: That is such a paltry listing of what a wife does. What 
 about counseling, caring for children, sustaining discouraged 
 husbands, and helping neighbors; that is a short list. 
 
 Student: Is there a possibility that the Bureau of the Census 
 could calculate what the average woman with a 'certain educa- 
 tional level earns in outside employment, and then use that figure 
 to impute the output of housewives? 
 
 Dr. Jones: That could be a possible approach if it is decided that 
 the accounts need to be changed, although this approach is so 
 general that the assumptions are very questionable. Your com- 
 ment also emphasizes one very serious problem that requires 
 immediate attention. Wives are discriminated against in calculat- 
 ing social security benefits because the tremendous contributions 
 of wives are not assigned a financial value. Under existing social 
 security laws, if a wife is divorced and she has a history of 
 earnings she will receive social security benefits when the various 
 age and other eligibility requirements have been fulfilled. Follow- 
 ing a divorce, each party simply retains the appropriate earnings 
 record accumulated. Yet, if the divorced woman is a nonworking 
 wife, what happens to her? She has no work record. She has no 
 accumulated earnings rate. In reality, she has worked at least as 
 hard as a woman who works outside of the home. This problem 
 creates a great social inequity when housewives are left unpro- 
 tected following a divorce. 
 
 Student: Would the revisionists advocate that the GNP begin to 
 include the monetary value of various intangible factors which 
 contribute to the total quality of life? 
 
 Dr. Jones: The critics basically argue that the current procedures 
 overemphasize the financial aspects of our system. 
 
 Student: You're getting out of hand. 
 
 Dr. Jones: Exactly. The examples are suggested to indicate the 
 complexity of the estimating problem. We too often assume that 
 statistics can be easily created by simply extrapolating averages 
 for large numbers of people. 
 
 Student: What about when a friend comes to you and asks for 
 help? You don't charge a counseling fee. You don't want to 
 measure all those things. 
 
 Student: I've been having difficulty following which side of the 
 argument you're taking. 
 
 Dr. Jones: That's the professorial function. I have attempted to 
 outline the dimensions of the issue. My personal opinion is that 
 the existing accounts should be improved but that they should 
 not adopt new definitions. The need for more quantitative and 
 qualitative information on the social aspects of our society can 
 and should move forward without destroying our existing 
 accounts. 
 
 Dr. Jones: That is a correct point. We don't attempt to measure 
 most of our social functions, do we? The same point applies to 
 the household output. 
 
 Student: What we are saying is: If you measure domestic help, a 
 woman who goes to another family's home and works from 9 in 
 the morning till 4 or 5 in the afternoon, why don't you measure 
 that for a regular housewife? 
 
 Dr. Jones: In the case of work outside of the home, a specific 
 financial value can be determined because wages are paid and 
 payroll taxes are collected by the Government. 
 
 Student: Let me ask exactly what we're measuring in the Gross 
 National Product. 
 
 Dr. Jones: That is the subject we started out with. The National 
 Economic Accounts are an estimate of the goods and services 
 produced, measured in financial values, for transactions occurring 
 within the market system. 
 
 Student: The revisionists are asking for a different system. This is 
 the way you introduced it. They seem to want something other 
 than the GNP. They must want a gross national wealth or 
 something. 
 
 Dr. Jones: 
 
 Welfare. 
 
 No, they want a MEW: a Measure of Economic 
 
 Student: Would that be in monetary terms? 
 
 Dr. Jones: It would have to be stated in some quantitative terms 
 and the existing common measurements use dollar values. For 
 example, what is the value of the television programs we watch? 
 We do not currently record an imputed value of television 
 programs in the GNP. We do include the advertising expenditures 
 used to support the programming. But the personal informational 
 and educational benefits of watching television are not incor- 
 porated in the existing GNP accounts. These are personal benefits 
 which have a value but which cannot be identified in dollar 
 terms. 
 
 45 
 
MEASURING NATIONAL WEALTH AND WELL-BEING 
 
 Another example of the difficulty of imputing financial 
 values involves the use of air-conditioning. The capital investment 
 needed to produce an air-conditioning unit is recorded. The 
 finished unit is reported as a change in business inventories. The 
 retail sale of the unit is recorded in the personal consumption 
 sector. The energy required to operate the unit is recorded in the 
 services account. But the personal enjoyment of being cooler does 
 not have an imputed financial value for entry into the GNP 
 accounts. 
 
 Student: I can identify the economic sacrifice I make to 
 purchase, maintain, and operate an air-conditioner over a period 
 of time. This is the value I ascribe to sitting around a room that's 
 72 degrees instead of 98 degrees. 
 
 Dr. Jones: And how did you arrive at that value? 
 
 Student: I arrived at it from the initial purchase price amortized 
 over the entire utilization of the machine plus the energy that I 
 had to pay for to operate the machine. 
 
 Dr. Jones: But all you've given me is one side of the account 
 which summarizes the outlays. I definitely agree that those values 
 are part of the National Economic Accounts. But how do you 
 value the benefits? Each of us would probably place a different 
 value on the enjoyment of air-conditioning. 
 
 Let me comment on one last category of activities which is 
 difficult to evaluate for GNP accounts. If something is merely a 
 necessity and it really doesn't improve the quality of life, should 
 it be included in the Gross National Product? For example, 
 should national defense outlays be included in the GNP? We 
 spend $95 billion a year on national security. Is this effort only a 
 "regrettable necessity" that should be deleted from the GNP? 
 
 Student: The purpose of national defense is safety. 
 
 Dr. Jones: How do you value that safety? You see, you're right 
 back in the box of using an imputed value for an output value. 
 Do you and I value defense the same way? 
 
 You could say that food should come out of the GNP 
 accounts because it's just a prevention of hunger. It doesn't add 
 anything. You could say that doctors' fees should come out 
 because they don't add anything by simply preventing illness. 
 This same argument could be made about all the so-called 
 "regrettable deficiencies." 
 
 We have a set of GNP accounts which are measured in 
 transaction terms because they are measurable values. You can 
 find out what a television set was sold for. You can determine the 
 value of the energy required to run an air-conditioner. You can 
 measure how much is being spent for national security. We need 
 these accounts stated in financial terms because such values serve 
 as a common denominator. This approach gets us away from 
 having everyone around this table making value judgments about 
 the relative value of air-conditioning using different techniques. It 
 is very useful to know what our economy has produced. 
 
 Decisionmaking 
 
 Of even greater importance, these National Economic 
 Accounts become the basis for the economic policy decisions 
 
 46 
 
 which are made every day. What should our monetary policy be? 
 Should we tighten, continue, or ease the monetary policy? What 
 should our Government spending programs be? Should they be 
 larger, the same, or less? 
 
 They also guide policy decisions affecting taxes by record- 
 ing the relative amounts committed to consumption and invest- 
 ment. The United States annually invests about 18 percent of its 
 Gross National Product in capital goods. The Japanese invest 
 about 38 percent. The Europeans invest anywhere from 25 to 35 
 percent. What does that difference mean? It may well mean a 
 great deal in terms of our future trade relationships and our 
 competitive position in the world. 
 
 We need to know how unemployment responds to Govern- 
 ment spending. If unemployment begins to rise, and it probably 
 will over the coming months, we will need a measure of current 
 economic activity to anticipate future employment conditions. 
 Over the last two quarters we have had a declining Gross National 
 Product measured in real terms after eliminating price increases. 
 When the real GNP is receding or growing at a rate of less than 4 
 percent, the economy is unable to absorb the approximately one 
 and three-quarters million people who enter the labor force each 
 year. 
 
 In other words, we need the Gross National Product to 
 measure (a) what we have done, (b) what we are doing at the 
 moment, and (c) how the pace of economic activity reacts to 
 various outside variables such as an oil embargo, a worldwide 
 decline in demand, or a crop difficulty. Such economic accounts 
 are really the fundamental basis for all of our economic planning. 
 
 Going back to your question, is the existing set of 
 economic accounts in any way inconsistent with knowing more 
 about the results of health programs, the quantity and quality of 
 the Nation's housing, or the number of people who have satisfied 
 their educational needs? Is there anything contradictory about 
 increasing the scope of our social statistics? Of course not. 
 
 My point is that it is not necessary to destroy or erode the 
 quality or purposes of the existing GNP accounts to achieve the 
 desirable goal of accumulating more information about the social 
 aspects of society. I know that officials of the Bureau of 
 Economic Analysis are constantly striving to improve the 
 accounts. In fact, I would argue strongly that we could make 
 greater gains by improving the current GNP accounts than we 
 would by diverting our attention to trying to value television, air- 
 conditioning or any of these other examples that we have 
 discussed briefly. We don't need a different conceptual Gross 
 National Product. Instead, we need a recognition that the GNP is 
 not a measure of the quality of life. It has never claimed to be 
 such a measure because it would represent only a partial indicator 
 of the total status of our society. There is no true measure of the 
 quality of life that I know of, because each of us has different 
 values. I would place a very high value on relaxation or the 
 chance to go home in the evening. Someone who works in a job 
 composed of constant repetition might be more interested in 
 gaining access to more interesting work. My wife might value air- 
 conditioning for our car. So, how can one measure the quality of 
 life? 
 
 What I want you to gain from this discussion is that the 
 Gross National Product is useful, in a very specific way, for 
 economic policymaking. But it clearly is not a measure of the 
 quality of life. I would strongly support the expansion of 
 programs seeking information on social issues as a basis for 
 making difficult decisions about welfare reform, the health 
 budget, educational programs, environmental standards, and the 
 
MEASURING NATIONAL WEALTH AND WELL-BEING 
 
 like. However, the basic GNP accounts should not be used for the 
 wrong purposes, nor should they be underrated as a measure of 
 economic activity. 
 
 Student: That seems perfectly logical. That makes me wonder 
 further, would the revisionists feel that all social planning, or 
 planning in regard to trade, finance, recessions, or whatever, is 
 distorted because somehow the GNP as it stands is inadequate? 
 
 which is significantly below the current target of 4 percent, will 
 moderate some of the strains in acquiring needed resources and 
 protecting our environment. This kind of information is ex- 
 tremely important in economic policy decisions. 
 
 Planning the Economy 
 
 Student: What is the outlook for inflation? 
 
 Dr. Jones: I don't think the attack would be on the basis that it's 
 a poor economic indicator. The criticism usually concerns the 
 narrow scope of the analysis. 
 
 Mr. Barabba: Let's say the Administration comes out and says 
 that they're going to set a policy of a 4-percent growth rate 
 and all these items mentioned. Then they go out and they meet 
 that rate and maintain it over a period of time. And they say, 
 "You see how great we are." Well, the revisionist would say, 
 "But, how great are you really? What have you done to the 
 environment? How happy are people?" 
 
 Dr. Jones: The classic example involves the United Kingdom. 
 About a year and a half ago they set a 5-percent annual growth 
 goal for their economy. They did not achieve that objective. In 
 fact, they are experiencing a negative economic growth rate in 
 real terms at this time. The GNP accounts would record that 
 result, but Government economists would probably argue that 
 outside factors have interfered and the shortfall is not the result 
 of poor policies. The social critic might go even further and claim 
 that the economic results are not an effective way to describe 
 social conditions in Great Britain. 
 
 Economists use the Gross National Product to indicate 
 current economic growth rates and the current mix of consump- 
 tion and investment. From these reports analysts can estimate the 
 probable course of future economic events, but they cannot 
 predict how the people will react to future developments. 
 
 We also need to know what the future holds. The Bureau of 
 the Census tells us that in the 1980's this Nation will have certain 
 work force patterns. They estimate that from 1958 to 1965, the 
 work force grew about 1.5 percent each year. From 1965 to 1978 
 it is expected to increase at an annual rate of 1.8 percent. This 
 small percentage shift has created severe employment strains. For 
 example, teenagers comprise approximately 10 percent of the 
 total work force, yet they account for over 25 percent of the 
 current unemployment. This particular shift in the labor force has 
 had a major impact on various manpower programs. 
 
 By the 1980's the labor force's annual rate of growth is 
 expected to be only 1.1 percent. The surge of teenagers resulting 
 from the postwar bulge of births will have been absorbed. 
 Fortunately, the 25- to 30-year age bracket of the labor force 
 tends to have much more stable work characteristics. Unemploy- 
 ment may be less of a problem in the 1980's, not because the 
 people will be more skilled, better educated, healthier, happier, or 
 any of the other characteristics measured by social indicators, but 
 because, in a relative sense, there will be fewer new workers that 
 will be entering the labor force. 
 
 Boiled down to the essential point, the economic system 
 concerns providing jobs for people. The system is supposed to be 
 for the people. During the 1980's the real GNP can increase at an 
 annual rate of only 3.2 percent and still absorb the new entrants 
 into the labor force. This reduced rate of economic expansion. 
 
 Dr. Jones: The United States has historically experienced a low 
 rate of inflation, averaging 2 percent during the 1900's. Severe 
 price disruptions have occurred as a result of wars, but the normal 
 peacetime rate has been very moderate. Beginning about 1965, 
 the pace of price increases started to accelerate, and eventually 
 crested in 1969 at an annual rate of just over 6 percent. In 
 response to the restrictive monetary and fiscal programs initiated 
 in 1969 (including the income surtax passed in June 1968), the 
 rate of price increases gradually receded. In 1972 the Consumer 
 Price Index rose 3.4 percent. Unfortunately, prices again turned 
 upward in August 1972, due to earlier fiscal and monetary 
 policies which caused the economy to become overheated. The 
 price indices literally exploded in 1973. There were many reasons 
 for the surge of prices: food shortages, the quadrupling of oil 
 prices, the convergence of worldwide demand pressures as most 
 industrial nations expanded rapidly, and the accummulation of 
 serious distortion created by 3 years of unfortunate wage and 
 price controls. 
 
 Over the last 12 months the GNP price deflator has in- 
 creased about 1 1 percent. This broad measure of inflation jumped 
 12.3 percent during the second quarter of 1974. The outlook is 
 that this intolerable rate should gradually decline to a single-digit 
 pace, but the improvement will be very difficult and painful. 
 
 With the notable exception of West Germany, where the 
 current rate of inflation is 7 percent, most European countries are 
 experiencing price increases at annual rates of 12 to 20 percent. 
 Japan is suffering an even higher inflation and in some countries 
 the increases are uncontrollable. 
 
 Student: Is the U.S. Government doing anything about this? Are 
 they trying to stop it? 
 
 Dr. Jones: Economic policies are focused on the two most 
 meaningful actions possible. The monetary policy has been 
 moderately restrictive for over a year. Over the last 3 years the 
 growth in the money supply (new currency and demand deposits 
 in the system) has grown at an annual rate of 7 percent. In 1973 
 the increase was limited to 6 percent and a target of 5 to 6 
 percent is now feasible and desirable. 
 
 The Administration is also striving to slow down the 
 momentum of the Federal budget outlays for fiscal year 1975 by 
 establishing a target of $300 billion. This target still represents a 
 massive increase of $32 billion over fiscal year 1974 and will 
 result in the fourteenth deficit in the last 15 years. Nevertheless, 
 achieving that goal would be a major step in regaining control 
 over Federal spending. 
 
 Beyond these two fundamental policies, the Congress and 
 Administration are always working on food, energy, regulatory, 
 antitrust, and hundreds of other specific programs which, 
 hopefully, will improve the efficiency of our economy and help 
 moderate the unacceptable price pressures we are currently 
 
 47 
 
MEASURING NATIONAL WEALTH AND WELL-BEING 
 
 experiencing. For example, policy decisions have been made to 
 release millions of areas of croplands that were previously held 
 out of production. After almost 40 years of artificially restricting 
 agricultural production, we are now striving to maximize output. 
 A major effort is going into the development of an effective 
 energy program. The price of oil will never drop back to the old 
 levels of $2 to S3 a barrel, but we can move ahead on developing 
 new energy resources. In short, the Government is attempting to 
 push a comprehensive package of economic policies which will 
 gradually improve the situation, but we should not have 
 unrealistic expectations that progress will be easy or quick. 
 
 Student: Won't these decisions ultimately amount to the old 
 question of priorities? 
 
 Dr. Jones: The entire subject of economics focuses on the push 
 and pull of setting priorities and allocating resources. 
 
 Student: Percentage wise, how have salaries gone up in terms of 
 the budget? 
 
 Dr. Jones: I don't know for the total Federal Government. I do 
 know that in the Department of Defense military personnel has 
 declined from 3 million people in 1969 to a current level of about 
 1.5 million, while the payroll costs have doubled. For the total 
 Federal Government, employment in 1969 totaled 2.8 million. 
 Last month the total was surprisingly 100,000 lower than 5 years 
 ago. However, State and local government employment increased 
 by 2 million to a level of 11.5 million as of July 1 974. Very few 
 critics recognize that Federal employment has declined while 
 State and local units have sharply increased their employee num- 
 bers. 
 
 Student: Is the U.S. economy dependent upon worldwide trade? 
 
 Dr. Jones: It is influenced by but not dependent upon it. 
 Merchandise exports have traditionally represented about 4 
 percent of the GNP. This year, exports will amount to about 7 
 percent of our GNP. In many European nations exports account 
 for 40 to 50 percent of their GNP totals. The United States is the 
 world's largest exporter and importer, but as a share of our total 
 economic activity, free trade is not as important as it is in most 
 other industrial nations. 
 
 Student: Is it possible to have a stable international economy 
 when inflation is so rampant? 
 
 Dr. Jones: The current levels of inflation are clearly disrupting 
 international economic stability. Part of this serious problem 
 involves the unprecedented jump in oil prices, but there are many 
 other causes. As the head of the U.S. delegation to the recent 
 meetings of the Economic Policy Committee in Paris, I empha- 
 sized the basic point that the United States has a major 
 responsibility to stabilize its domestic economy as a basis for 
 international economic progress. The situation is somewhat 
 analogous to being in a canoe with an elephant. No matter how 
 friendly or well mannered he is, if he makes a mistake, a disaster 
 occurs. The U.S. economy is so large and important that we have 
 a unique responsibility to provide worldwide economic leadership 
 by pursuing responsible fiscal and monetary policies. 
 
 Student: How are your data systems being exchanged? 
 
 Dr. Jones: We exchange economic information at meetings of the 
 Organization for Economic Cooperation and Development 
 (OECD), the United Nations, the International Monetary Fund, 
 and of many other international institutions. 
 
 Mr. Barabba: You might go into the reconciliation of trade 
 statistics that the Bureau of the Census worked out with Canada. 
 
 Dr. Jones: That exercise is a good example. For many years 
 U.S. -Canadian relationships have been exacerbated by disputes 
 about the trade balance figures. Since each nation keeps its own 
 record of both exports and imports, the disparities between the 
 large trade deficit with Canada reported by the United States and 
 the much smaller surplus figure reported by Canada have been a 
 source of confusion and occasional irritation. For example, in 
 1973, the United States originally reported a deficit of $2.6 
 billion in our trading account with Canada for calendar year 
 1972. Canada reported a small surplus of about $0.6 billion in 
 our trading account with Canada. As part of the reconciliation 
 process, the U.S. Bureau of the Census and its Canadian 
 counterpart discovered many problems of definition and interpre- 
 tation. The cooperative effort of these two excellent agencies 
 clarified these problems and resulted in an effective reconciliation 
 of the figures. After adjusting for definitional problems, mistakes, 
 and missing documents, the reconciled figures for 1972 indicate 
 that the United States really experienced a deficit of $1.2 billion 
 in its trading relationship with Canada in 1972. 
 
 Statistics should not dominate international negotiations, 
 but it is important to get our facts straight before beginning to 
 discuss policy issues. The Bureau of the Census is now talking to 
 many other countries about similar reconciliation projects. This is 
 the basic thrust of my message here today. Government officials 
 try to establish monetary and fiscal policies on the basis of 
 meaningful National Economic Accounts. 
 
 Student: On a long-term basis, if there is not a decrease in the 
 price of oil, will the current economic policies have much impact 
 in terms of righting the economy and bringing down long-term 
 interest rates? 
 
 Dr. Jones: Fortunately, the United States is about 85-percent 
 self-sufficient as to sources of energy. Nevertheless, the dis- 
 tortions caused by the sudden oil price increases have seriously 
 disrupted our domestic economy. The effects have been devastat- 
 ing in many other countries. 
 
 Student: Won't the difficulties in other nations hurt our 
 economy if they can't buy our products? 
 
 Dr. Jones: Economic distortions in other nations will restrict our 
 exports in some situations; however, our food exports are 
 experiencing increasing demand pressures and our traditional 
 exports of high-technology products will continue. It is true that 
 U.S. merchandise exports are currently running at an annual rate 
 of about $90 billion, which is well above the 1973 total of $70 
 billion and the 1972 figure of $49 billion. Our shipments of 
 agricultural products were about $20 billion last year. We 
 certainly do not want to see trade wars or a serious erosion of our 
 
 48 
 
MEASURING NATIONAL WEALTH AND WELL-BEING 
 
 exporting activities, even though the total figure accounts for 
 only about 7 percent of our GNP. 
 
 Even more serious are the disruptions of financial flows. 
 The Arab oil-producing nations are going to run surpluses in 1974 
 of $50 to $60 billion. They can import maybe $5 or $10 billion 
 more, but that would still leave approximately $40 to $50 billion 
 for investment. The other part of the issue is that advanced 
 OECD nations will have trading deficits of $40 to $50 billion. 
 The surpluses of oil-producing nations must somehow be "re- 
 cycled" to finance the imports of oil-consuming countries. 
 
 Even more important than the "recycling" of oil earnings is 
 the question of the price of oil. If oil prices remain at such high 
 levels, the United States must turn to developing energy 
 alternatives. This Nation has the technological capabilities to 
 develop its coal, oil shale, and nuclear facilities, and to move 
 ahead on solar, water, and wind-power solutions to the energy 
 crisis. We can develop these long-term options, but the immediate 
 problems for the lesser developed nations are extremely serious 
 because they just can't buy oil with their limited resources. It is 
 also a difficult situation for Europe, because they import about 
 two-thirds of their energy needs, and for the Japanese who must 
 buy about 99 percent of their energy resources. The current 
 distortions caused by the sudden increases in oil prices are 
 probably the single most difficult problem in the international 
 economy at this time. Fortunately, trade is continuing, and we 
 have avoided the international monetary crises that we ex- 
 perienced in 1969 through 1971. 
 
 Student: If you do all that you've talked about and implied, 
 won't that carry a higher unemployment rate? 
 
 Dr. Jones: If business cycle fluctuations can be moderated by 
 monetary and fiscal policies, then unemployment will fluctuate 
 within a more narrow range. Unfortunately, economic policies 
 have tended to add to either boom or recession because the 
 programs are changed so frequently. This adjustment process is 
 difficult because events tend to lag about 1 year behind policy 
 changes. By the time the policy adjustments affect the economy, 
 events have changed so that the shift exaggerates the underlying 
 trends. A major reason for having more stable economic policy is 
 to avoid the disruptive swings which create uncertainty in the 
 private sector and lead to possible unemployment and lost 
 investment. During the first quarter of 1973, the real GNP (price 
 increases removed) was increasing at a seasonally adjusted rate of 
 9.7 percent. Incidentally, 44 percent of the American people felt 
 we were in a recession at that time, even though we were growing 
 at twice our sustainable annual growth rate. The economy 
 obviously could not continue to operate at that intense pace, so 
 fiscal and monetary policies were used to gradually cool down the 
 situation. 
 
 Student: Wouldn't you want to show welfare spending in the 
 GNP so you could forecast that a bit better? 
 
 Dr. Jones: Welfare expenditures are included in the Government 
 spending accounts. A large portion of the outlays are reported in 
 the State and local government sector. 
 
 Student: A couple of times you've mentioned taxes. But, in 
 terms of reducing Federal spending, isn't there also the alternative 
 of raising taxes? 
 
 Dr. Jones: Right. 
 
 Student: Is this policy avoided because the raising of taxes is so 
 politically bad? 
 
 Dr. Jones: Changes in the tax code require legislation. History 
 suggests that it is extremely difficult to increase taxes because it 
 is hard to convince the American public that such drastic action is 
 required to pay for all the spending programs demanded. 
 
 Student: There are a number of commentators saying that 
 Government's preoccupation with Watergate has aggravated our 
 inflation situation. Could you comment on that? 
 
 Dr. Jones: I think our inflation is infinitely more complex than 
 that simplistic explanation. If it has had an impact I think it 
 would be minimal. The uncertainties created may have curtailed 
 consumer spending. The stock market apparently has also been 
 negatively influenced by recent political events. As far as 
 significantly influencing the underlying rate of inflation, I can 
 give you all kinds of reasons that are much more important than 
 the psychological impact of Watergate. 
 
 Student: What can we expect in the future with this new 
 economic policy? Will the inflation rate go up for awhile, then 
 curve down, and then turn back up? Will it ever level off? 
 
 Dr. Jones: The wholesale price index for July was 3.7 percent. If 
 that rate were to continue for 12 months, it would result in a 
 50-percent gain in wholesale prices in 1 year. What happened was 
 that prices of industrial commodities such as steel, copper, 
 plastics, chemicals, etc., were influenced by the sharp increases in 
 oil prices. The distortions created by oil price gains will be 
 winding their way through the entire system. 
 
 For 4 months in a row farm prices dropped sharply. Then 
 in June and July the weather destroyed many crops. Wholesale 
 food prices in July went up 6.8 percent, which is the equivalent 
 of an annual increase of about 84 percent. Prior to July, there 
 were cross trends with industrial raw material prices pushing up- 
 ward and wholesale farm prices declining. In July, rising farm 
 prices reinforced the upward trend of other raw material costs. 
 Farm output this year will not be as good as originally expected, 
 
 but it will be better than last year. 
 
 One encouraging aspect of the inflation outlook is that 
 
 energy prices are not going to quadruple as they have in recent 
 months. In fact, there seems to be some evidence of an 
 oversupply of crude petroleum, amounting to approximately 2 
 million barrels a day, around the world. Unfortunately, interna- 
 tional oil decisions involve complex political issues. Oil-producing 
 nations can partially offset pressures to reduce prices by cutting 
 back their output. Certainly, the oil prices are no longer going up, 
 and they may be cracking a little bit. 
 
 Based on the outlook for food and energy, we expect 
 inflation will slowly and painfully moderate. It may sound as if 
 I'm talking about a different world on the same day the wholesale 
 price index went up 3.7 percent. But if you step back and look at 
 the last 4 or 5 months, considering what you see in the economy 
 and what you anticipate concerning food and energy, I think it's 
 fair to say that inflation will gradually moderate if a tough 
 economic program is sustained. By the end of the year, I don't 
 think we will still be suffering double-digit inflation. But I could 
 well be disappointed. In the United States an inflation rate of 2 
 
 49 
 
MEASURING NATIONAL WEALTH AND WELL-BEING 
 
 percent has been our historical average. Today, if we could get it 
 down to 4 percent, that would be a Nobel Prize winning perform- 
 ance. This achievement will not occur over the next few months. 
 Nevertheless, we must strive to reduce inflation in order to avoid 
 the social distortions that are already developing. Inflation is 
 particularly hard on families with lower incomes. Unfortunately, 
 a disproportionate share of our low-income families are from 
 minorities that are still vulnerable to discrimination. People who 
 are on fixed incomes and the elderly who must watch their 
 savings eroded by galloping inflation while they wonder whether 
 or not they will outlive their savings are also tragically affected by 
 current price increases. Inflation is really a very vicious thing in a 
 social sense. 
 
 Student: I was surprised that you said that the economists are 
 now in general agreement about general policies. 
 
 Dr. Jones: They really are. It seems so obvious that we just have 
 to require more fiscal restraint. 
 
 Student: You mean economists in the private sector as well? 
 
 Dr. Jones: Yes, and in the public sector. 
 
 Student: I was particularly surprised because everything I've read 
 in the newspapers in the last few weeks has been on how the 
 economists are in a quandary as to how we got into this mess and 
 how we're going to get out of it. Nobody agrees on which is the 
 right course of action. 
 
 Dr. Jones: I believe that there is a consensus about fiscal policy 
 and the need to improve the efficiency of the system. There is 
 some disagreement about monetary policy, but even here the 
 range of views is relatively narrow. It's not that we don't know 
 what to do, it's just doing it. 
 
 Mr. Barabba: Would Congressmen also accept fiscal restraints? 
 
 Dr. Jones: Most Congressmen agree about the need to better 
 control the total outlays. There are individual disputes, of course, 
 about which programs should be expanded and which ones 
 should be cut back, but there is a growing recognition that the 
 upward momentum must be controlled. In fiscal year 1974, 
 Federal outlays totaled $268 billion. The original fiscal year 1975 
 projections anticipated outlays of $305 billion. That upward 
 surge of $37 billion in a single year represents an increase of 13.8 
 percent in an economy which has had zero real GNP growth over 
 the last four quarters and which has had only an 8.3-percent 
 expansion of GNP measured in inflated prices. That kind of 
 budget growth distorts the economy. Furthermore, fiscal year 
 1975 Government spending could very easily have risen to $310 
 or S312 billion if strong action had not blunted the upward surge 
 of spending in this election year. 
 
 Fifty-four senators recently sent the President a letter 
 calling for a reduction of $10 billion in this year's Federal 
 spending, suggesting that such action would cut spending from 
 S305 to S295 billion At that time the budget was on its way to 
 S310 or $312 billion. It is interesting to note the first vote of 
 these 54 senators. When the agriculture expenditure bill calling 
 for several billion dollars was voted upon it included an increment 
 of S250 million over the spending target. Of the 54 senators, 49 
 voted that day, and 33 of those voted for the increase, while 17 
 
 50 
 
 of them voted to add another $600 million. In a related action, 
 46 of the 49 voted for a program to guarantee loans to livestock 
 producers. Last week they cut the Department of Transportation 
 budget 3 percent across the board, however, they didn't mention 
 that they exempted the highway trust funds, which account for 
 90 percent of the total. 
 
 Even though everybody seems to support a program of 
 fiscal responsibility, the actions of governments must support the 
 current rhetoric. Unfortunately, most governments do not have 
 much credibility. Fiscal restraint must be actually demonstrated 
 before most people will believe that the historical pattern of 
 constant deficits will be reversed. In fiscal year 1969, approxi- 
 mately $7 billion was cut out of the Federal budget and strong 
 monetary policy actions were adopted. The 6 percent rate of 
 inflation that existed in 1969 was slowly reduced to a level of 3 
 percent by 1972. Progress can be made. 
 
 Student: But didn't the Consumer Price Index pretty well 
 skyrocket during 1973? 
 
 Dr. Jones: The Consumer Price Index rose 8.8 percent during 
 1973 It is interesting to note that personal incomes increased 
 10.6 percent last year. 
 
 Student: What are the chances of Congress actually doing 
 something about the Federal budget? 
 
 Dr. Jones: When one talks about cutting the budget, you really 
 envision ranking all the projects. Let's say there are 60,000 
 Government projects. The economic approach would be to rank 
 every one of those projects and then proceed to eliminate those 
 at the bottom of the list until the necessary savings were 
 accumulated. Do you know what you would probably find when 
 you tried to eliminate the project numbered 59,990? You would 
 discover three powerful groups: first, a very vociferous group of 
 beneficiaries; second, a very antagonistic Congressional com- 
 mittee with oversight responsibilities; and third, a very strong and 
 protective Executive Office agency responsible for running that 
 particular program. It may sound easy to find likely projects for 
 elimination, totaling at least $5 billion in a $305 billion budget. 
 In reality, every attempt to reduce spending creates a strong 
 backlash from all three of these groups. Nevertheless, a reduction 
 of $5 billion of outlays during fiscal year 1975 would represent a 
 cut of only 1.6 percent of the total Federal budget. It is hard to 
 believe that a reduction of 1.6 percent could destroy any 
 Government agency. Unfortunately, this crude across-the-board 
 reduction cannot be used because most existing programs are 
 required by legislation. This problem is particularly meaningful 
 for those of us interested in the improvement of Government 
 statistical programs. The desirability of such improvement is 
 obvious to everyone who works with statistics. But we need to 
 remember that such programs are expensive and they must 
 compete against thousands of other priorities. 
 
 Mr. Barabba: Approximately $350 million will be spent on 
 information-gathering activities in 1974. Did you indicate that 
 the United States might be returning to wage and price controls? 
 
 Dr. Jones: No, I hope we don't. However, I do think that there 
 will be extensive pressure for such action. 
 
 Student: Aren't there other suggestions for cutting down 
 inflation? 
 
MEASURING NATIONAL WEALTH AND WELL-BEING 
 
 Dr. Jones: There are at least four important approaches. First, 
 we need more fiscal responsibility. Second, monetary policies 
 should be more stable. Third, the Federal Government should act 
 decisively to eliminate or reduce the inefficiencies and distortions 
 caused by its own policies. Finally, some analysts believe that 
 wage and price controls should be used. 
 
 Based on my analysis of the results of such controls in 
 Europe and the United States, I believe that wage and price 
 controls are not a meaningful way to control inflationary 
 pressures. In fact, I believe that they create serious distortions. 
 First, there are artificial incentives to export more domestic 
 goods whenever uncontrolled foreign prices are higher. Second, 
 controls delay or curtail capital investment which eventually 
 contributes to supply shortages which increase price pressures. 
 Third, there is a tendency to eliminate many low-priced items 
 from the product mix. Fourth, there is a tendency for wage and 
 price decisions to move upward to the permitted ceiling to avoid 
 being frozen at a lower level. 
 
 Student: What if you just froze wages instead of prices? 
 
 Dr. Jones: You would have a general strike. 
 
 Student: Economically, it wouldn't be feasible anyway, would 
 it? 
 
 Dr. Jones: I do not think so. 
 
 Student: How about consumer boycotts? 
 
 Dr. Jones: The consumers did boycott beef when it became too 
 high-priced. Consumers should not waste money during good 
 times or bad times. They should buy things that are lower priced. 
 What we don't want is further erosion of personal consumption 
 which has been flat for over a year. 
 
 Another serious problem involves the costs of energy. To 
 avoid further capacity shortages, the current utility rates may 
 well have to be raised to encourage capital investment and to 
 provide the necessary capital. Thirteen utilities have already 
 cancelled construction projects, most of which involved nuclear 
 facilities. The current loss of capital investment projects will 
 affect utility operations many years from now. The current 
 situation is a great paradox. We obviously want to curtail 
 inflation, but it is probable that State regulatory utility commis- 
 sioners will have to allow rate increases to generate the necessary 
 capital investment. The reaction of consumers to higher utility 
 rates is naturally going to be very negative. 
 
 Student: I think one reason why consumers feel the way they do 
 is that there seems to be, real or not, exorbitant profits for all 
 kinds of companies, be they oil, dog food or peanut butter 
 companies. When you stop eating meat and go to peanut butter, 
 and then that price goes up on you, it is very discouraging. 
 Companies all seem to say the same thing: "Unless I can make a 
 certain profit margin, I'm not going to do it. Go elsewhere, or 
 suffer." I agree with you fully that the American consumer is 
 spoiled rotten and is terribly wasteful. We're beginning to suffer 
 now and we're probably going to have to adopt different living 
 habits, but who is in a position to deal with the profit motive? Or 
 in many cases, the excess profit motive? If I have the correct 
 figures, some oil companies, during the first two quarters, 
 doubled their profits. Nobody is saying that they don't need 
 
 money to continue to look for oil wells, but what do you do with 
 a recalcitrant industry that says, unless I make a certain amount 
 of profits, I will not invest? They simply claim that the "business 
 of business" is to make a profit and if they don't, they will shift 
 investments elsewhere. What do you do with that kind of an 
 attitude? 
 
 Dr. Jones: Point number 1: Corporate profits as a share of 
 national income have been declining over the last 10 to 20 years. 
 They now have a lower share of national income than in 1937. 
 Point number 2: Over any extended period of time, oil company 
 profits have been about average compared with all U.S. industries. 
 Steel companies are also currently reporting sharp percentage 
 increases in profits in the last few quarters, but they have con- 
 sistently been near the bottom of the list each year when indus- 
 tries are ranked according to profitability. Steel companies now 
 face massive capital investment requirements, including very large 
 commitments for pollution abatement programs. Over the last 
 decade steel companies invested $17 billion for new plants and 
 equipment, but they did not increase capacity because the funds 
 were required for modernization and replacement projects and 
 the pollution abatement requirements. 
 
 As to the oil companies, many of them are reporting very 
 large percentage gains in profitability this year. Part of this jump 
 results from the statistical impact of comparing a large gain 
 during the first quarter of 1973 to a depressed profit report for 
 the first quarter of 1972. Also, more than 50 percent of the 1975 
 profit improvement is accounted for by earnings on foreign oil 
 operations. The companies do not believe that these results are 
 representative of the long-term outlook. Unless a sustainable 
 profit incentive is available, the oil companies will remain 
 reluctant to commit the massive capital amounts required. Many 
 basic industries have not added the necessary capital investments 
 needed to avoid shortages 
 
 Student: Don't companies receive certain advantages, such as tax 
 breaks, for investing abroad? Are we that helpless as a Govern- 
 ment or as a people, that we can often take certain measures to 
 control the consumer but cannot control industry? 
 
 Dr. Jones: We can nationalize our industries, like the British steel 
 industry or the European railroad system. But I think that you 
 will find that a nationalized industry becomes a tax drain rather 
 than a taxpayer. I strongly support the vigorous enforcement of 
 antitrust laws, but I do think that our people will have to 
 recognize that corporate profit incentives are lagging behind as 
 the basic motivation required for domestic investment. 
 
 Student: Is that why the Government is stepping in and 
 guaranteeing loans? 
 
 Dr. Jones: The Government has guaranteed loans for the 
 Lockheed Corporation and certain defense contractors. A pro- 
 gram to guarantee loans to livestock producers was also estab- 
 lished recently, but such examples are relatively rare. 
 
 Student: There was an article in the New York Times that the 
 lack of competition and the Government support for Lockheed 
 and the livestock men is what's causing prices to remain high, and 
 if you returned to a more competitive economy, you would have 
 a drop in prices. 
 
 Dr. Jones: I would certainly support increased competition. 
 Increasing productivity is also an excellent anti-inflation ap- 
 
 51 
 
MEASURING NATIONAL WEALTH AND WELL-BEING 
 
 proach. In the case of the Lockheed loan guarantee, I was 
 philosophically opposed to the decision. But there were some 
 strong practical arguments in favor of providing support to 
 prevent a corporate failure. It was claimed that 60,000 jobs 
 would be saved. When the loan guarantee decision was made, 
 California had an 8-percent unemployment rate. It is my 
 understanding thdt Lockheed has not tapped the Federal Treas- 
 ury for any funds. Nevertheless, when one borrower is granted a 
 preferential advantage in arranging a private loan, other borrowers 
 are placed in a disadvantageous position. The issue becomes 
 particularly pertinent when some mortgage borrowers have to pay 
 interest rates of 10 percent or more. The accumulated impact of 
 Federal deficits also disrupts the financial markets because the 
 Treasury must borrow large amounts of capital to finance the 
 excessive spending. Because the Federal Government is the largest 
 and most powerful borrower in the capital market, the individual 
 mortgage loan is usually at the bottom of the list, just behind 
 small businessmen. The impact of Federal deficit financing has 
 not been emphasized enough to alert the general public as to its 
 impact on interest rates and credit availability. These are just a 
 few examples of the complexity of establishing effective Govern 
 ment policies, particularly programs to reduce inflation. The basic 
 approach is to rely on monetary and fiscal policies along with 
 specific actions to improve productivity, but it will take a 
 considerable period of time to make significant progress in 
 reducing inflation. Unfortunately, interest rates are not going to 
 come down until inflation pressures are moderated, and, until 
 interest rates decline, many serious problems in our business and 
 housing industries will persist. 
 
 Student: One major alteration in the Federal budget might be a 
 Federal health insurance program; do you think that adding $60 
 billion to the budget would be inflationary? 
 
 Dr. Jones: In 1969, the Council of Economic Advisors prepared 
 a study which attempted to project the probable Gross National 
 Product over the next 5 years and the specific claims that would 
 be made against that output. Various assumptions were made 
 about Government spending, business investments in plants and 
 equipment, residential construction programs, and the traditional 
 patterns of personal consumption which result in the spending of 
 91 percent of our personal disposable income on ourselves. The 
 basic conclusion of that study was that the entire potential out- 
 put was already committed to existing claims, even though the 
 study assumed that a relatively rapid rate of economic growth 
 would be sustained throughout the entire period. The obvious 
 significance of that conclusion is that society will have to either 
 curtail or eliminate something from its current portfolio of spend- 
 ing programs if new projects and different priorities are to be 
 accepted. 
 
 Student: Is it a matter so much of new priorities as a way of 
 redistributing? 
 
 Dr. Jones: It is actually a problem of restraining the upward 
 momentum of the total budget and realigning priorities. 
 
 Student: No, I was going back to the health question. The 
 Kennedy-Mills sort of proposal. 
 
 Dr. Jones: I don't know that much about the mechanics of that 
 particular bill, but I thought the thrust of your question was: If it 
 winds up with more spending, will that be inflationary? The 
 answer to that is, yes. 
 
 Student: My point is that it isn't spending, per se, that creates 
 inflation. I think we have to take into account, in terms of in- 
 flation, the ways it's being spent, both in the payment of taxes 
 and personal expenditures. 
 
 Dr. Jones: I assume it would be somewhat analogous to Medicare 
 and Medicaid. From 1965 to 1969, the daily cost of a hospital 
 room increased 69.7 percent. Part of that remarkable increase was 
 caused by the surge of demand for medical services and facilities 
 without a corresponding increase in the number of doctors, 
 nurses, hospital rooms, support personnel, etc. Economists would 
 refer to this situation as an imbalance resulting from excess 
 demand pressing against an inelastic supply. Again, I don't 
 know that much about the National Health Program. I would 
 emphasize that it is naive to think we can simply increase 
 spending without eliminating other existing claims. Many people 
 believe that large cuts in defense outlays would provide the 
 necessary funds to pay for expanded social programs. However, it 
 should be recognized that defense spending has already been 
 sharply curtailed, particularly in a "real" sense when the eroding 
 effects of inflation are considered. Social program spending is 
 already well above defense spending. Further cutbacks in defense 
 outlays will be difficult. 
 
 When I speak at university campuses, students often suggest 
 that resources could be released by reducing business investment. 
 This is a rather naive suggestion because such capital outlays are 
 the fundamental variable in producing the future GNP. The 
 historical advantage of the U.S. economy has been the technologi- 
 cal competitiveness created by investment. Unfortunately, we 
 now lag far behind almost every other industrial nation in the 
 share of the national output being committed to current invest- 
 ments required for future growth. America has emphasized cur- 
 rent consumption while needed capital investment has suffered. It 
 has been estimated that the United States will need to commit 
 over $3 trillion to new capital investment over the next decade. 
 This an astounding estimate, but it can be accomplished if the 
 rate of annual outlays is somewhat increased and the Federal 
 Government does not usurp the capital to finance its deficits. 
 
 It is also possible to challenge the assumption that 
 consumers will, or that they should, continue to spend 91 percent 
 of their disposable personal income on themselves. I personally 
 believe that this allocation of resources to current consumption is 
 most short-sighted. It should be noted that the economist does 
 not say that it is right to emphasize personal consumption so 
 much. It is rather discouraging that we spend more money on 
 beauty and barber shops than we do for religion, welfare, 
 research, and educational activities. The economist must explain 
 the repercussions of continuing the historical consumption 
 trends. If society wants to spend less on existing types of personal 
 consumption and more on other needs with longer term benefits, 
 I would applaud that decision. But you have to go out and 
 convince 213 million American people to do that. 
 
 Congress is now attempting to consider budget priority 
 issues more carefully by creating a joint committee to evaluate 
 Government spending. This new bill which was signed on July 24 
 gives the Executive Office authority to specify individual spend- 
 ing programs that should be delayed or eliminated entirely. The 
 Congress must approve such recommendations before the adjust- 
 ments can be made. If the electorate could really communicate to 
 their Congressmen and the Executive Office that they want 
 Federal spending to be more controlled, then some fundamental 
 changes might occur. 
 
 52 
 
PROFILESOF SMALL URBAN AREA CHANGE 
 
 Richard Hanel, Director of Public Relations, R. L. Polk Company 
 
 "You cannot judge a city in terms of its total complex. You've got to take it apart 
 and look at the pieces. Further, no two cities are alike, even though some of the totals 
 may look alike. Cities are just like people. They have their ins and outs, ups and downs, 
 and pluses and minuses-some react quickly while others move slowly, and they react in 
 their fingers and their toes. They are virtually unpredictable when it comes to small area 
 year-to-year change." Rjchard Rane| 
 
 Mr. Barabba: The speaker this morning is Mr. Dick Hanel of the 
 R. L. Polk Company, one of the oldest companies in the 
 information business in the world, and one of the few that is still 
 privately owned. Since 1967 I'd been hearing about Dick Hanel 
 and finally met him after I became Director of the Census 
 Bureau. 
 
 He came to the Bureau one day and gave a presentation 
 which was of interest to the Bureau. He showed our professionals 
 what Polk was doing with statistics based on the information 
 generated from data in its city directories. Polk wasn't calling 
 these statistics social indicators; yet, they were clearly indicators 
 of certain phenomena within communities. Polk has a unique 
 capacity for providing current data because they're out there 
 surveying most metropolitan areas once every year. 
 
 Mr. Hanel: As Mr. Barabba said, the Polk Company started 
 publishing city directories over 100 years ago, and directories 
 remain, to this day, the biggest part of our business. We also 
 publish bank directories, motor vehicle statistics, and so on. So, 
 basically, we, too, are in the information business. 
 
 This morning, I'm going to be talking about that part of our 
 business called urban statistics which is a byproduct of our city 
 directory operations. In order to set the stage, let me describe the 
 city directory. 
 
 It's a reference book which costs the customer from $35 to 
 $155, depending on its size and the area covered. Directories 
 consist of three sections: An alphabetical listing of people, a 
 classified business section, and a listing of addresses in street and 
 house number sequence. All together we publish about 1,200 
 different directories containing information for about 7,000 cities 
 and towns. Most are published annually. 
 
 The information in city directories comes from a door-to- 
 door canvass. We don't ask for a lot of information, and 
 everything we ask at a household— with one exception, number of 
 children— is printed in the directories. Our questions are very 
 simple. What is the name of the head of the household and what 
 is his wife's name— the latter question immediately sets up marital 
 status. What are the names, job titles, and employers of everyone 
 18 and over in that household? We ask for the number of children 
 in each household, if the home is rented or owned, and the phone 
 number. We also collect information on type and location of the 
 business and professional establishments in the community. 
 
 That's the data base we're going to discuss. From this data 
 base, using computers, we compile small-area inventories of the 
 housing, the people, and the businesses in cities. More importantly, 
 we report on the year-to-year changes that are taking place. My 
 
 main mission this morning is to introduce you to that dimension 
 of change. 
 
 It is our conviction that change is really the name of the 
 game, not just change in city totals, but change in small 
 areas— neighborhoods. The rates and geographic patterns of 
 change are so vitally important in understanding what's going on 
 in cities that we call our product "profiles of change." 
 
 Our statistical operations are frequently compared to the 
 Census Bureau, and some people have even said that we're 
 competitors. We don't look at it that way at all. We're only 
 competitors in the very, very, limited sense that we do some of 
 the same things. Both Polk and the Census Bureau publish 
 statistical information about cities, and people frequently com- 
 pare our information with the census counts of 1970. Actually, 
 there are more differences in what we do than similarities, and I 
 would like to describe a few of those differences for you. 
 
 First of all, the Census Bureau collects its information 
 primarily for the purpose of compiling statistical data. On the 
 other hand, we collect our information primarily to publish city 
 directories and to make a legitimate profit on the sale of these 
 directories and the advertising they carry. So, the statistical work 
 that we do is a recent byproduct of a commercial enterprise that 
 has been going on for a hundred years. 
 
 Secondly, the Census Bureau investigates deeply in the 
 course of some of its probes. On the other hand, we work with 
 only a very few items which are essentially the public face of the 
 individual; this is the same information we have been collecting 
 and publishing for many years. 
 
 The third big difference between us and the Census Bureau 
 is in the economics of the finished product. If the Bureau is asked 
 to go out and do a special census of a city today, they will charge 
 about 50 cents per person to make a survey from which they will 
 give the city census tract population counts, with breakdowns by 
 sex and age, race and ethnic group, and so on. That 50 cents per 
 person comes to about $1.50 a household if you assume three 
 persons per household. By virtue of the fact that our statistical 
 information is the byproduct of the ongoing city directory 
 operation which has already paid for itself, everything that we are 
 going to talk about today— dealing with housing, people, and 
 businesses, inventory information plus the change calculated from 
 two surveys— can be provided to a city for less than 15 cents a 
 household. This price puts it very much in the ballpark in terms 
 of the kind of current viable statistics that a city can afford, 
 although our information would be limited to the information in 
 the city directory. 
 
 This is a good example of the point that Mr. Barabba has 
 
 53 
 
PROFILESOF SMALL URBAN AREA CHANGE 
 
 stressed, low cost byproduct availability of data is going to 
 become a very, very important factor in the dissemination and 
 use of vital information. 
 
 People ask us what kind of a sample we get. Well, our 
 objective is to get as close to a 100-percent complete coverage as 
 is possible and, in fact, we are generally successful in making 
 contact with about 90 percent of all the occupied housing units 
 in a city. The contact rate will go higher in certain more stable 
 neighborhoods; it may be lower in poorer neighborhoods and 
 those with a fast turnover, but we've never seen it below about 65 
 percent. Those who are expert in statistics and sampling tell us 
 that this should be a very representative sample. 
 
 I'm not going to belabor this background anymore, except 
 to say that, nationwide, our directory people knock on the doors 
 of 27 million housing units every year. We have statistics on the 
 central cities and most of the suburbs in 209 out of 253 SMSA's 
 (standard metropolitan statistical areas), including some big cities 
 like Detroit, Houston, Boston, and Washington. We don't make 
 directories for Los Angeles, New York, Chicago, Philadelphia, or 
 Baltimore. The economics of doing city directories in some of 
 these large cities just does not pan out. I think that sets the stage 
 for you. 
 
 From now on we're going to show you some of the things 
 we've learned about the "profiles of change" in cities. 
 
 Possibly one of the most interesting things about cities is 
 the rapid rate of household movement. We measure it in terms of 
 the number and percentage of housing units that have had at least 
 one family move in or out in the time between two directory 
 canvasses. 
 
 As we measure it, there are five components to this rate of 
 change. First, how many housing units which we found occupied 
 at the time of the first canvass are now vacant? Second, how 
 many went the other way, from vacant to occupied? Third, how 
 many newly constructed housing units have been occupied? 
 Fourth, how many previously occupied units have been vacated 
 and demolished? Fifth— and usually about half of the total 
 action— how many housing units have seen at least one family 
 move out and another move in from the time of one canvass to 
 another? 
 
 In most cities, we have found that from 25 to 35 percent of 
 all housing units have had at least one family move in or out in a 
 year's time. We've even found a few cases where the city-wide 
 movement rates were from 45 to 55 percent! And when you get 
 down to the census tract or neighborhood level, the change rate 
 frequently runs into the 60 to 80 percent, and even higher. 
 
 In some cities and neighborhoods, rapid move rates have 
 meant a drastic, radical change in family counts and demography. 
 In other cases, families move, but the demography profile of an 
 area stays pretty much the same. There's no depending on it. 
 
 This is why, as one of the few generalizations we permit 
 ourselves, we say that you cannot judge a city in terms of its total 
 complex. You've got to take it apart and look at the pieces. 
 Further, no two cities are alike, even though some of the totals 
 may look alike. Cities are just like people. They have their ins and 
 outs, ups and downs, and pluses and minuses— some react quickly 
 while others move slowly, and they react in their fingers and their 
 toes. They are virtually unpredictable when it comes to small 
 area, year-to-year change. 
 
 For these reasons, the other generalization we feel 
 somewhat safe in making is that no one providing statistics is by 
 that fact alone qualified to understand what is happening in a 
 
 city- It takes a terrific dose of local knowledge to interpret the 
 numbers and to assess their meaning and importance, and we try 
 our best to keep our role in this whole business to that of 
 submitting a report to our client cities. 
 
 How do we report the data? We almost always work in the 
 dimension of the census tract, which is the base point most 
 people prefer as a building block. We furnish 22 statistical tables 
 and over 500 counts, percentages, and ratios for each census tract, 
 all generated from that small data base that we talked about. One 
 of the reasons for the amount of numbers is that we're dealing in 
 the dimension of location. 
 
 The second thing that we do for the city is to gather key 
 information in terms of maps. Location and patterns in the data 
 are vital, and the best way to see them is with maps. We print 
 over 300 computer maps for each city, dealing with as many 
 counts and percentages and covering all subjects. Today, you're 
 going to see maps, for example, dealing with vacancies in our 
 example city. Anybody who wants to can probe the subject of 
 vacancies using maps. 
 
 Now, on the subject of social indicators, we're doing 
 something crude, simple, but extremely effective. For each small 
 area, such as a census tract, we pull together on one page of the 
 report 56 key counts and percentages. Then we rank each census 
 tract in the city on these counts and percentages, for example, 
 percentage of heads of household with no jobs. The tract in the 
 city of 100 census tracts with the highest percentage of the 
 jobless gets the rank of 1. The tract with the lowest percentage of 
 jobless household heads gets the rank of 100. And we do this 
 with the computer for each of the 56 counts in each one of the 
 tracts. 
 
 This very simple rank order is proving to be something that 
 people can understand, and it is turning out to be the common 
 denominator for many things, providing rudimentary social 
 indicators in terms most people can comprehend. Factor analysis 
 is great, regressions are fine, but people don't understand some 
 data conclusions, and don't know how to read others. But people 
 can understand simple rank order. 
 
 Now I want to talk about the way we account for change. 
 We rack up the information in terms of inventories and rank 
 orders. Basically, what we do is take this year's computer tape for 
 a city and using the computer, almost literally lay it on top of last 
 year's tape. We make an address-by-address, family-by-family 
 comparison of what we found this year versus what we found last 
 year in terms of the components of the changes that have taken 
 place. We're not just looking at the difference in the counts of 
 vacancies or the net difference in households as of two points in 
 time. We're looking at the input and the output, the flow, the 
 velocity of that change and its components. That's the important 
 thing. 
 
 Information Use 
 
 How are cities using information on change? First of all, 
 much of the current legislation in Congress with respect to cities 
 is toward block grants as distinct from categorical grants. (Note: 
 the Community Development Act has since been signed into law.) 
 The old way of doing things allocated so much money for model 
 cities, so much for sewers or urban renewal, in other words, to 
 specific programs. This discouraged local initiative in the use of 
 money except within that framework. 
 
 54 
 
PROFILES OF SMALL URBAN AREA CHANGE 
 
 The thrust of the Community Development Act is to give 
 the city a block grant, which is one, big, fat check. You've heard 
 it called revenue sharing. It is up to the cities to determine where 
 to put their priorities, money, attention, and effort. Cities are 
 using available information to set their own priorities and sort 
 these problems out quantitatively. Cincinnati has been doing a lot 
 of work in the area of neighborhood priorities. 
 
 The second thing, of course, is that once you have a feel for 
 change, you can start to develop programs and strategies that take 
 into account the forces at work. Again, for 3 years Cincinnati has 
 been developing and refining a neighborhood-by-neighborhood 
 housing strategy. 
 
 Thirdly, and we don't want to overstate its present 
 capability, there is evaluation. There can be some fairly straight- 
 forward evaluation. Substandard housing can be removed, but 
 you have to be able to measure your relocation needs and 
 capabilities, and the results of the moves in terms of both housing 
 and people. It's not an easy thing to do unless you have some 
 pretty good numbers. Norfolk has some excellent results to show 
 for its efforts in housing and social programs. 
 
 Among the most important elements of planned com- 
 munity growth is communications— being able to talk to the 
 people involved in, or affected by, these programs and to explain 
 what's going on and where people or programs stand. Milwaukee, 
 Wise, one of the best examples, has a very active neighborhood 
 development program. Its been getting into all kinds of problems, 
 social, housing, health, and so on, by using numbers to explain to 
 neighborhoods, "This is where you are." People stopped specu- 
 lating on how bad the situation was, and found themselves 
 instead, and very quickly, with some hard numbers, trying to do 
 something to change the objectionable ones. 
 
 I'm going to show you three things today, first, some of our 
 maps to give you an idea of the sweep of things and how change 
 can affect one town. 
 
 At the end of that presentation you'll see how quickly 
 small-area change can affect events at the census tract level. I'll 
 try to demonstrate how some statistically unimportant numbers 
 can be extremely meaningful in the art of understanding 
 neighborhoods and anticipating problems before the problems 
 overwhelm the occupants and the decisionmakers. 
 
 Finally, you'll see a time series for one city which 
 illustrates the movement and force of some of the basic trends 
 that are at work these days. 
 
 You'll recall, a few minutes ago, we were talking about the 
 velocity of change, the percentage of housing units in cities 
 involved in some kind of change of occupancy. The numbers on 
 the following maps are printed by the computer. Over the 
 printout we lay a transparency which shows, in this case, the 
 outline of the census tracts. The whole thing is simply reduced to 
 8Y2" x 1 1" size; and we give each city about 300 different maps. 
 In about 2 hours you can run through 40 key maps for a city, 
 mark them up to show the highs and lows and patterns— the sort 
 of thing we're showing you here— and get an overview of what's 
 going on in the city. 
 
 Here's the city of Milwaukee (map 1). Of Milwaukee's 
 230,000 housing units, 27 percent had at least one change of 
 occupancy. The shades of gray indicate the percentage of all 
 housing units in a single area that had at least one family move in 
 or out. 
 
 The pattern is fairly common. Usually, the high move rates 
 are found in two places, around the central business dbtrict, 
 
 which you see at the east central part of town, and where new 
 housing developments have been constructed, in this case in the 
 northwest. 
 
 Here's the city of Cincinnati (map 2). Cincinnati's move 
 rate is nearly 10 percentage points higher than Milwaukee's. 
 Again we find the areas with high move rates, 50 percent or more, 
 around the central business district. The more stable areas are 
 out toward the periphery of town. But note that it's generally a 
 more busy looking pattern. 
 
 Here's a complete change-San Antonio (map 3). There is 
 the same 37 percent overall movement rate that we saw in 
 Cincinnati. This exemplifies what happens when a city builds a 
 ring road around itself— lot of new construction around the 
 periphery of the city, with the usual high move rate in the center 
 of town. 
 
 Next is the granddaddy of them all, Houston, Tex. (map 4). 
 The move rate is 44 percent with still a different pattern. 
 Concentration of change is in the business district, the center of 
 town. But, here we see a spoke-of-the-wheel pattern reflecting 
 growth along the major arteries which radiate from the center of 
 Houston. 
 
 We have seen city-wide movement rates as high as 50 to 55 
 percent particularly in Colorado Springs, El Paso and Corpus 
 Christi. 
 
 Now we're going to take a closer look at a midwestern city 
 that we'll call Midtown, U.S.A. (map 5). Here we're back to a 
 28-percent move rate, with the usual concentration of action 
 around the central business district (CBD). 
 
 Now, what's going on beneath this 28-percent move rate, 
 which is well within normal range? Midtown has about 60,000 
 housing units so the 654 new ones equal a 1 -percent increase 
 added by new construction concentrated in the eastern part of 
 town. 
 
 While that was happening, about 400 housing units were 
 lost due to demolitions, mainly in the center of town, around the 
 CBD. Watch that area because that's where we're going to see a 
 lot of action. 
 
 We also measure change in the housing unit count which 
 does not involve new construction, demolitions or even a change 
 in vacancies. This is your single to multiple family residence 
 conversions— two families living in a house where there was one 
 before with no identifiable change in the vacancy situation. Note 
 that this is Midtown's biggest change of all, a net increase of 943 
 housing units. But, again in the CBD. 
 
 Now, here is an example of how fast change can occur. In 
 the previous year, instead of realizing an increase of 943 new 
 units, the city suffered a loss of 429 units in precisely the same 
 part of town. 
 
 When you add it all up— new construction, demolitions, 
 other change— Midtown had a net gain of 1,200 housing units in 
 two hunks (map 6). In the eastern part it was primarily new 
 construction. In the central part of the city there was mainly 
 non-physical change. 
 
 The next component is vacancies. It is not unusual to find 
 that, of the more than 600 newly constructed housing units, 139 
 were vacant when we made our canvass. Of the more than 400 
 housing units lost to demolitions, about 200 of them, mostly in 
 center city, had previously been vacant. 
 
 But here is the important one. Within existing housing there 
 was a net increase of 437 vacancies (map 7), concentrated in the 
 same central part of town where we found the increase in housing 
 units from sources other than new construction. This is the sort 
 
 55 
 
PROFILES OF SMALL URBAN AREA CHANGE 
 
 of thing that you sometimes find in areas undergoing rapid 
 transition— an increase in vacancies in existing stock and doubling 
 up going on at the same time. Meanwhile, vacancies were down in 
 the southwestern part of the city, reflecting absorption of new 
 construction from the previous year. 
 
 Add it all up for Midtown and there's a net increase of 400 
 vacancies with the important rise in the center of town. 
 
 You may recall that there were 1,200 new housing units. 
 Offset that with 400 more vacancies and you have a net increase 
 of 800 in the number of occupied housing units or households, 
 with major small-area change showing gains only (map 8). 
 
 And again, just to show you how quickly things can shift, 
 the previous year the change in households was plus 291, with a 
 pattern that shows both plus and minus areas, including abdica- 
 tion of a lot of housing right in the center of the city. 
 
 The next thing we find is that despite an increase of 800 
 households, Midtown lost nearly 200 people as a result of 
 households moving in and out. This reflects the general trend 
 toward smaller households, including increases in the number of 
 one-person households. This kind of population loss is very 
 measurable, and usually takes effect in the areas surrounding the 
 CBD, the older parts of the city. 
 
 Here we have a good example. In one census tract, 578 
 households were moving in at an average size of 2.22 persons per 
 household. There were 576 moving out, virtually a draw in 
 household count. But the average size of the families moving out 
 was 2.45. That difference between 2.22 incoming and 2.45 
 outgoing was enough to account for a loss of 134 people in one 
 tract, even though the number of households remained almost 
 precisely the same. We find this frequently. 
 
 In Midtown's demography, first, we find 20 percent 
 one-person households, pretty much what the Census Bureau 
 finds. But against a gain of 800 households total, the gain of 441 
 one-person households is far more than proportionate. 
 
 Next, households with children number 42 percent. The 
 interesting thing is the rather large concentration of households 
 with kids in the center of town. Usually you find these families 
 closer to the periphery. Then, against a city-wide gain of 800 
 households in total, there were only 156 more families with 
 children. 
 
 The theme of your meeting is social indicators, and here is 
 one that is getting a lot of attention, female heads of house with 
 children. In Midtown the rate is about 7 percent, a normal figure 
 in most cities. More important, we find a gain of 225 such 
 households as a result of movement. This is interesting because it 
 exceeds the total gain of households with children. 
 
 Retired heads of households number 21 percent, slightly 
 higher than usual. The important thing, though, in Midtown is the 
 loss of 262 retired heads of house in the central part of town. 
 
 For jobless heads of house, our counts and rates exceed the 
 unemployment rate as the Bureau of Labor Statistics defines it. 
 Our figures also include people who are unemployed but not 
 necessarily currently looking for a job; people on welfare; 
 divorcees and others, but not students or retirees who are not in 
 the labor force. The 10 percent that we found in Midtown is 
 pretty much in line with the nationwide rate of the Census 
 Bureau. The important thing in Midtown is, of 800 more 
 households, there is an increase of 700 households headed by 
 people jobless or not in the labor force (map 9). We are finding 
 this sort of thing in most large cities. In Midtown, it's important 
 that much of this change was concentrated just to the east of the 
 city. The poor area is the highly mobile area. 
 
 However, we don't suggest that people use our data as the 
 last word. First of all, it will have changed already by the time 
 they use it. Secondly, perfectly precise data need not be used for 
 planning most major programs or strategies. 
 
 The important and unique thing in our data is that it shows 
 short-range, small-area change. This is the vital dimension. Change 
 data plus the inventory counts make it possible to see the main 
 patterns and trends and identify areas that may need further 
 investigation. 
 
 In Midtown, homeowners number 66 percent (map 10). 
 That rate nationally fluctuates from 45 to 65 or 70 percent. 
 Homeowners are normally found near the periphery of the city. 
 
 The important thing is the loss in this city of 500 
 homeowners versus a net gain of 500 households, in the area east 
 of the high vacancy neighborhood and in the same vicinity where 
 we noticed the loss of retirees. 
 
 By contrast, we find a gain of 1 ,300 units rented. The main 
 shift from households owned to rented occurred in the central 
 part of the city, a frequent but not universal characteristic. In a 
 deteriorating area people in need can't afford to buy a vacated 
 house, while it just isn't attractive to someone who can afford to 
 buy. So it becomes a rental unit. 
 
 Student: Do you know what happens to that owner? Do they 
 become a renter in the outer periphery of town? 
 
 Mr. Hanel: We don't know today, but the potential is in the data 
 to find this out. There are several things that can be done that we 
 have not yet had a chance to do. One of them is this very 
 important question— the so-called chain of moves. What happens? 
 Where do people go? Why? People talk a lot about the "chain of 
 moves," and there's a lot of speculation about "filtering" and 
 "moving up," etc. But, relatively little factual work has been 
 done on that subject, particularly involving large areas and 
 numbers of cases. 
 
 One of the important things that we can furnish is an 
 address-by-address tape showing precisely what we found the 
 current year and if there has been a change from the previous 
 year. From that tape you can identify the new movers— you find 
 the addresses into which someone moved. You call these people 
 on the telephone and say, "Where did you come from?" You get 
 their previous address and you call the people who are now there, 
 and you start tracking backward. This, we think, is going to be 
 one of the important byproducts of this whole business. 
 
 For the purpose of this discussion, the example we just gave 
 also raises the subject of being able to do selective low cost 
 samples from the detail files. 
 
 Some time ago people at the Department of Housing and 
 Urban Development said, "Your vacancy data are great but we'd 
 like to know two more things; how many bedrooms are there in 
 these units and what is the sale or rental price?" So they took 
 four cities and, using our vacancy listing, went to the local 
 telephone company and said, "Here are some addresses. Have any 
 of these since been occupied and do they now have telephones?" 
 Most of them had been occupied in 2 or 3 months— the turnover 
 of vacancies is fast. They took the phone numbers and for 50 
 cents a household they made calls, got the information on the 
 number of rooms, the sale or rental price, and other information, 
 and tabbed it up. 
 
 Take the journey to work. We have the name of the 
 employer in the household record and this can be cross tabbed 
 with the business file to set up a matrix at the tract level. We're 
 
 56 
 
PROFILES OF SMALL URBAN AREA CHANGE 
 
 starting now to have the time and the facilities to get into these 
 sort of things. 
 
 The next subject is income. People in the cities always want 
 us to ask questions on income. We don't even try for two reasons: 
 Income is a very, very difficult thing to determine, and we know 
 darn well that we would have doors slammed in our face all over 
 the place if we asked income questions, which certainly would be 
 in the area of personal privacy. However, Mr. Barabba and his 
 friends at the Census Bureau came along and gave us the kinds of 
 answers we need. They are now publishing a series of data, based 
 on a nationwide sample survey, showing total household money 
 income according to the occupation of the head of the 
 household. As a sample, this series takes into account the fact 
 that a household with a jobless head might have two or three 
 workers and sources of income, whereas a household headed by 
 someone who is retired possibly has only social security benefits. 
 The latest available Census Bureau numbers show average 
 total household money income was about $11,300 nationwide. 
 Doctors, it turned out, had the highest average household 
 incomes, about $34,000. We converted doctor dollars into an 
 index number by dividing $34,000 by $11,300, and used the 
 answer (303 percent) as the index. In this way, we compute index 
 numbers for 21 categories of households according to the 
 breadwinner's occupation. 
 
 We do a number of things with the index numbers. First, 
 we compute a very simple arithmetic mean or average index for 
 each census tract and the city as a whole, in all 21 categories. In 
 Midtown, we found that the citywide index reflecting the 
 occupational mix of the heads of house was 102, or slightly 
 higher than the average of the national sample. 
 
 On map 1 1 we show those census tracts in Midtown where 
 the index number was 112 or more and where it was 92 or 
 less— roughly the high and low income areas. 
 
 Second, we compute the average income for movers-in, and 
 for movers-out and compare the net difference between the two. 
 Using these differences you can get a very clear view of where the 
 major changes in the high counts tend to stretch to the east. 
 
 Also, we calculate the available housing stock. According to 
 HUD, a minimum of 1.5 percent of the total owner-occupied 
 housing stock should be vacant and available for sale to take care 
 of normal turnover. Despite Midtown's 7-percent overall vacancy 
 rate, there are only 300 such housing units for sale. These, too, 
 are just east of the low income trouble area that is seeing so much 
 action (map 12). 
 
 The other side of the coin, vacancies for rent, is another 
 story (map 13). There are about 800 housing units available for 
 rent over and above 4 percent of the renter stock and they are 
 smack in the middle of the low income area. 
 
 We break down our vacancies into two kinds. First, we have 
 two-canvass vacancies, which are housing units vacant at the time 
 of each of two annual surveys. There's no guarantee that someone 
 didn't move in or out during that period, but usually there have 
 been no tenants. Where we find high rates of such vacancies, 
 you're apt to have deterioration, dilapidation, or abandonment. 
 In Midtown, the tracts with 50 or more such vacancies are 
 centered in the low income areas. They generally go together. 
 
 The counterpart of two-canvass vacancies are those housing 
 units either occupied at the time of the previous canvass and now 
 vacant, or newly constructed and vacant— current-year vacancies. 
 Thi'.- should be your viable housing stock. I n Midtown we find out 
 that there are about 2,700 current-year vacancies, which is about 
 4 or 5 percent of the total housing stock. All but one of the low 
 income tracts have high counts of current-year vacancies. This is a 
 
 good example of the importance of breaking down the totals to 
 find out what's going on. 
 
 Those of you dealing daily with urban affairs know one of 
 the big problems is housing. And it's going to be a bigger problem 
 as time goes on, considering high construction costs, high interest 
 rates, high everything. 
 
 Right now, and particularly in center city Midtown, 
 housing construction is almost stagnant. Pressing questions for 
 the next few years are how much viable housing stock is in the 
 city, and how do we make the best use of it. Midtown has a 
 7-percent total vacancy rate and by normal standards some 
 people would say the town is in pretty good shape on vacancies. 
 But, discounting the two-canvass vacancies, the fact most of the 
 current-year vacancies were located exactly where nobody 
 wanted to live shoots the 7-percent surplus! This is occupational 
 mix taking place. It's turning out to be a very sensitive kind of 
 rating of neighborhood stability. 
 
 The income index numbers are proving to be important in 
 understanding residential vacancies. Midtown has nearly a 
 7-percent vacancy rate and that's mid-range. We're finding 
 everything from 3 or 4 percent up to 13 percent. The city-wide 
 7-percent vacancy rate looks like this (map 14): 12 census tracts 
 have 10 percent or more vacancies, ranging as high as 29 percent, 
 and 10 of the 12 are in the low income area. 
 
 We also map some aggregates of the income index points to 
 determine concentrations of household income in Midtown. 
 Next, we see where changes in total household income took place 
 as a result of changes in the number of households, or average 
 household income. City-wide, Midtown's $10 million increase in 
 household income is concentrated primarily in areas with higher 
 household counts. It's important that no tracts show significant 
 decreases. 
 
 These diagrams of aggregate income lead us into a look at 
 Midtown's professional and business community. You can't 
 neglect the supporting businesses and services when you try to 
 understand the forces at work. Midtown's business and residential 
 communities remain pretty much segregated. While most of the 
 commercial vacancies are centered in the CBD they're spilling 
 over into some of these areas where the income is concentrated 
 and growing. 
 
 More important, the number of professional and business 
 establishments declined during the year, influencing some of the 
 areas where the number of people, households, and the aggregate 
 disposable income were increasing. More cities are beginning to 
 consider these factors, and talk to their business community in 
 terms of the need and opportunity for supportive services. The 
 summary map includes detailed counts of inventory and change 
 for a number of types of business. However, this area needs a lot 
 more work and understanding. 
 
 Earlier we talked about rank ordering of all tracts in terms 
 of each factor. Of those 56 factors, we have found 8 that turned 
 out to be extremely sensitive and useful as indicators of the 
 general health and well-being of a neighborhood. First is the 
 percentage of housing units that have had a change of occupancy. 
 The second factor is the percentage of housing units that are 
 vacant, excluding new construction. Third is the percentage of 
 two-canvass vacancies. Fourth is the percentage of business 
 vacancies. Fifth is the average income index with the lowest given 
 the rank of one. Sixth is the percentage of jobless heads of 
 households. Seventh is the percentage of female heads of house 
 with kids. And last is the percentage of one-person households. 
 Rankings based on these eight factors have helped identify and 
 quantify a city's priorities. 
 
 57 
 
PROFILES OF SMALL URBAN AREA CHANGE 
 
 The conclusion of our rankings is that a lot of the change in 
 Midtown is taking place in the areas where big problems currently 
 exist, and generally, the change is not good. 
 
 The second major thing I promised was to show you how 
 quickly small-area change can affect and characterize a census 
 tract. Here we're getting into some of these so-called "non-statis- 
 tically-sound" kinds of techniques. But, it's the kind of approach 
 that illustrates the importance of local knowledge. 
 
 We're going to be talking about three adjacent census tracts 
 in Midtown (table 1); one (tract 32) is in the southeast corner of 
 the critical area that we talked about. Each is heavily populated, 
 ranging from 1 ,200 to 1 ,800 households. 
 
 Table 1. COMPARISON OF SMALL-AREA CHANGE 
 
 Characteristics 
 
 City 
 
 Tract 
 32 
 
 Tract 
 33 
 
 Tract 
 34 
 
 
 61,156 
 
 1,7 93 
 
 1,159 
 
 1,714 
 
 
 102 
 
 92 
 
 111 
 
 118 
 
 Current Percentages 
 
 
 
 
 
 Housing units with change 
 
 
 
 
 
 of occupants 
 
 28 
 10 
 
 37 
 23 
 
 25 
 
 11 
 
 15 
 
 
 3 
 
 Female household heads 
 
 
 
 
 
 with children 
 
 7 
 
 7 
 
 52 
 
 20 
 10 
 90 
 
 13 
 
 5 
 
 54 
 
 3 
 
 Vacancies 
 
 2 
 
 Total 
 
 23 
 
 Net Change in Counts 
 
 
 
 
 
 Households 
 
 (MA) 
 (NA) 
 
 49 
 
 75 
 
 1 
 
 18 
 
 -10 
 
 
 -5 
 
 Female household heads 
 
 
 
 
 
 with children 
 
 (NA) 
 (NA) 
 
 51 
 110 
 
 22 
 36 
 
 7 
 
 Renters 
 
 9 
 
 
 
 Now we do something that makes a statistician shudder. We 
 just add up the percentage of housing units with change of 
 occupancy, jobless heads of households, female heads with 
 children and units vacant. For our first tract, the sum of the four 
 is 90 percent; next door, it is 54 percent; farther away, 23 
 percent. City-wide, the total is 52 percent. 
 
 There you have the current status. Three tracts— one of 
 them is obviously in trouble; another, by the current numbers 
 about average; and a third much better than average. 
 
 How about trends? Tract 32 shows a net increase of 49 
 households, mainly from doubling up. Against that, an increase of 
 75 jobless household heads, an increase of 51 female bread- 
 winners with kids, and a shift from owner to renter results in a 
 net increase of 1 10 renters. There's a tract already in trouble and 
 cntinuing to head downward as evidenced by the lower income 
 and impermanence. 
 
 In tract 33, our average tract, the number of households 
 basically didn't change, but jobless heads of households increased 
 by 18, the female heads of house with children increased by 22, 
 and the renters grew by 36. Read by the numbers— later 
 confirmed by local knowledge— here's a tract that's average in 
 terms of where it stands today, but starting to hit the skids. 
 
 Seeing tract 34 your antennae begin to quiver a little bit. 
 This is one that's still well off, but vacancies increased by 10 
 households. Heads of households who were jobless decreased by 5 
 but, female heads of house with kids were up by 7, and an 
 increase of 2 from owner to renter in the face of all that. A mixed 
 bag of small changes adding up to a neighborhood that they're 
 
 going to want to watch because it could be in the way of the 
 wave of deterioration. 
 
 Table 2 shows you some time series for the city of Norfolk, 
 Va. I pulled together this time series because I thought it 
 interesting to see what happens in a well-managed city in terms of 
 trends. They started with 260,000 people in 1970 (excluding 
 military living on base) and 3 years later were down to 248,000, 
 while the number of households stayed about the same. The 
 explanation is the family size which declined steadily from 3.11 
 to 2.95. 
 
 Look at the number of housing units, down from 88,800 to 
 87,200, not a lot of net change to be sure. But look at the 
 amount of new construction— 1,500 units in 1971, 1,700 in 1972 
 and 2,000 in 1973. 
 
 Table 2. PROFILES OF CHANGE COUNTS 
 NORFOLK, VIRGINIA 
 
 Characteristics 
 
 1970 
 
 1971 
 
 1972 
 
 1973 
 
 
 260,500 
 88,800 
 
 252,600 
 87,300 
 
 249,200 
 86,700 
 
 247,800 
 
 
 87,200 
 
 Units added by new 
 
 
 
 
 
 
 (NA) 
 
 1,500 
 
 1,700 
 
 2,000 
 
 Percent of units with 
 
 
 
 
 
 (NA) 
 
 41.3 
 
 37 .9 
 
 36.4 
 
 
 83,600 
 3.11 
 
 82,000 
 3.08 
 
 83,000 
 3.00 
 
 83,900 
 
 Persons per household... 
 
 2.95 
 
 Vacancies., number.. 
 
 5,200 
 
 5,300 
 
 3,700 
 
 3,300 
 
 percent . . 
 
 5.7 
 
 6.0 
 
 4.3 
 
 3.8 
 
 Two- canvass 
 
 
 
 
 
 
 V NA) 
 
 1,500 
 
 1,200 
 
 900 
 
 percent . . 
 
 (NA) 
 
 1.7 
 
 1.3 
 
 1.0 
 
 Current-year vacancies.. 
 
 (NA) 
 
 3,800 
 
 2,500 
 
 2,400 
 
 Current-year vacancies 
 
 
 
 
 
 vs. desired minimum.... 
 
 NA) 
 
 1,240 
 
 44 
 
 -38 
 
 Current Percentages 
 
 
 
 
 
 Household with 
 
 
 
 
 
 
 48.1 
 
 48.1 
 
 46.8 
 
 46.1 
 
 Female heads with 
 
 
 
 8.4 
 
 8.5 
 
 8.6 
 
 9.4 
 
 Retired heads of 
 
 
 
 12.0 
 
 12.1 
 
 13.1 
 
 14.4 
 
 Jobless heads of 
 
 
 
 11.0 
 13.6 
 
 10.9 
 13.6 
 
 11 .1 
 14.9 
 
 12 .0 
 
 One-person households... 
 
 15.7 
 
 
 53.6 
 
 53.4 
 
 53.6 
 
 53.7 
 
 
 
 Student: That could also indicate a shift, destroying older houses, 
 and things like that. 
 
 Mr. Hanel: Exactly. And if you look at the next line you will see 
 one of the results-and this is one of the best signs in the whole 
 business. By 1973, their total vacancies were down to 3.8 
 percent, which is extremely low, because they have removed a 
 number of unusable vacancies. Two-canvass vacancies are steadily 
 down. Current-year vacancies are below minimum levels. They 
 cleared out the substandard housing in the old section of town. 
 So there's a very effective measurement of the results of an 
 aggressive kind of clearance and rebuilding program. They turned 
 their new housing into occupied housing, while getting rid of the 
 old stuff that was just sitting there. 
 
 On the other hand, by any standard, housing is extremely 
 
 58 
 
PROFILES OF SMALL URBAN AREA CHANGE 
 
 tight in the city, and in today's situation that could be considered 
 poor planning. But again, where you have few vacancies, you 
 generally have less problems of other kinds. 
 
 Mr. Barabba: A good point to make here is that when one says 
 something is an indicator of something good, we have to ask, in 
 whose eyes is it really good? That's one of the problems with 
 social indicators. 
 
 Mr. Hanel: A point was made that these eight indicators are 
 negative, and they are to the extent they were designed to try to 
 pinpoint the areas which might require priority attention from 
 the standpoint of problems. The other side of the coin is, what 
 would be the positive indicators in a city? Is it low ranking on the 
 same eight factors or a whole new set? 
 
 We get a lot of feel by looking at reports and maps, but I'd 
 hesitate to interpret the numbers for a city without a lot of local 
 background and knowledge. You can't avoid pointing out certain 
 things that stand out from what you learn to expect. But even the 
 level of vacancy rates, for instance, must be read against move 
 rates, local programs and attitudes regarding land use, and 
 clearance. Interpretation needs a heavy dose of local 
 understanding, and this is a different dimension from reporting. 
 
 Student: In learning to interpret the data and extrapolate from 
 what you have, do you find a political problem in addressing 
 people within a city? Do they like you wandering in the door and 
 saying, "Look, the place is going to hell in a handbasket. You'd 
 better do something!" Or do you have to nurse them around to 
 the viewpoint of understanding? 
 
 Mr. Hanel: It depends on the people involved. You can run the 
 gamut from the absolutely enthusiastic top man, to the apathetic, 
 to the "don't tell me about problems" types. The big question 
 mark in this whole thing and the question we run into frequently 
 is, "O.K., so we know all this. Now, what do we do?" I'm not 
 sure anybody knows that answer. But I'm sure of one thing— the 
 first step the "doers" are taking is to find out what the situation 
 is and what forces are at work. 
 
 Mr. Barabba: Let me add a couple of things. One is that within 
 the Bureau we have established what we refer to as the Census 
 Use Study. And we have, throughout the country, including here 
 in Los Angeles, research going on. In Indianapolis, we have what 
 we call the Uses Program. Anytime anybody in the metropolitan 
 government requires information, they are allowed to go into the 
 local Census office. Our contribution is quite extensive but the 
 payoff is that we're recording every request, the kind of 
 information provided, and the actual result related to that 
 information. So we're trying to learn more about how informa- 
 tion is used. 
 
 We then started a series of seminars called GBF/DIME 
 (Geographic Base Files and Dual Independent Map Encoding) 
 Workshops, and we're trying to move more and more into 
 simulated decision formats that allow people easy access to the 
 information for decisionmaking. It's a very difficult task because 
 you have to make it general in order that people can learn from 
 it. On the other hand, if you get it too theoretical, you run into 
 the same problem of having to go to school all over again to 
 understand. It's a kind of balancing act which I think is evolving 
 and we're getting more and more support now. The New 
 
 Federalism has probably contributed more towards our ability to 
 move our resources in this direction, because we can make the 
 case that governments have to give people more than money to 
 implement their ideas. This can be done by giving them the 
 wherewithal! to make better decisions, and showing how data can 
 be used to monitor and evaluate. 
 
 Mr. Hanel: We've done something along this line which is called a 
 User's Guide, which introduces the available information in easy 
 stages. We tell the clients how to mark the maps, what to look for 
 and what the benchmarks are. 
 
 The point is that our information package is just plain 
 overwhelming to a lot of people when they first get it. Imagine all 
 the paper— 22 multipage reports and 300 maps. Somebody takes a 
 look at that and says, "What do we do now?" The easiest thing to 
 do is put it on the shelf and forget it. And yet, it's all useful 
 reference and working material. We're trying to lick this problem. 
 The simpler you can make the material the better, but the 
 problem is it's hard to make things simple. We'd like to pare 
 down our package, but the deleted portion turns out to be 
 important information that somebody wants. 
 
 So we're always doing, as Mr. Barabba said, a balancing act. 
 The real problem is that of publicity. What you're seeing now is 
 our attempt to popularize this information program and get key 
 people involved. 
 
 Mr. Barabba: In these seminars that I mentioned, we try to 
 design a technical workshop where people, who would use the 
 information as part of their daily routine, will actually experience 
 the data files. We also have a 1-day presentation for their 
 supervisors where we really put on all the bells and whistles. In a 
 sense, what the Bureau is trying to do is to create a push and a 
 pull through the entire system. It's a very difficult task. But every 
 once in a while we are successful. We find somebody changed 
 their way of doing things. 
 
 Mr. Hanel: I couldn't help smiling while Mr. Barabba was talking 
 because, you see, he's giving information away, while I have to 
 sell it— however modest the price! 
 
 Another important point is the interpretation of the data. 
 Something which looks bad may actually be good. A loss of 
 population is not necessarily bad. A concentration of female 
 heads of house with kids is not necessarily negative, if they're 
 moving into housing intended for them. You have to accommo- 
 date these people and the real question is whether the particular 
 program is working the way you planned? 
 
 Let's finish up with our friends in Norfolk and look at the 
 trend in some of the household characteristics. Households with 
 children are down a couple of percentage points. Statistically it 
 might not mean that much, except it seems to be an underlying 
 trend in many central cities, and something the school people 
 have to contend with. Meanwhile, the percentage of female heads 
 of house with kids is increasing. That contrast is important, and 
 calls for a look at geographic distribution and patterns. 
 
 In Norfolk, like many other older cities, there is definitely 
 an upward trend in the percentage of retired heads of household 
 in the central part of town. The number of jobless heads is fairly 
 constant in Norfolk, which is a strong plus, particularly when you 
 see that the small-area geography is also stable. One-person 
 households are very definitely on the increase, something that 
 you're going to spot in city after city. In Norfolk, there was little 
 change in the high percentage of renters. This city really has 
 
 59 
 
PROFILES OF SMALL URBAN AREA CHANGE 
 
 grabbed hold of things. The people understand their town, they 
 are truly monitoring and managing it. 
 
 Let's get back to indicators. There's a lot of talk by 
 scientists and statisticians about indicators. The disappointing 
 thing, to me, is that some of those same people are going the 
 farthest underground and away from the real world. They're 
 hiding behind studies that conclude, "We don't dare talk about 
 something bad because that will make the situation worse." I 
 think we must make some fairly straightforward, easy-to-under- 
 stand, shirt-sleeve use of these types of numbers, such as the rank 
 order ratings. Let's experiment, test, and verify with sophisticated 
 analysis as needed, but always keep the end result in a common 
 denominator that people can easily understand. 
 
 There are all kinds of meaningful subjects that ought to be 
 worked on. Cincinnati is looking at the relationships between 
 some of these indicators and crime and fire. It turns out that 
 there's some real coincidence between the rates of crime and fires 
 and vacancies. This suggests that when you can find a single 
 strong indicator, in this case, vacancies. Then you can earmark 
 the places that deserve some special attention. 
 
 Student: In Los Angeles we have parts of this city where 
 dwellings have been demolished. The people who were living 
 there before. . .who knows where they're living now? 
 
 Mr. Hanel: To make a determination as to where the people went 
 would require another step beyond where we've gone, but the 
 resources are there to do it. You're asking if you do something 
 that displaces someone, what happens to him and what is the 
 rolling effect on others? Not by name, but certainly by numbers, 
 we can see where these people went. You might find that the 
 unemployment in some areas reworked through urban renewal is 
 very low; in other cases it might be naturally high because of the 
 presence of low-income housing. 
 
 We've seen some evidence in Norfolk of this sort of thing. 
 They cleared out some substandard housing and the people who 
 had lived there moved on. Now, there is evidence most of these 
 folks didn't leave town. The next question is, specifically, where 
 did these people move, and more important, what do they think 
 about* their new neighborhoods? How do you get at that one? 
 You take the tape that says, "Here are our new movers. These are 
 the people who have relocated within the last year." Pull out a 
 list of movers with the characteristic you're concerned about and 
 talk to them. 
 
 Mr. Barabba: Granted, we have some displacement. But, how do 
 you find out if we're doing the right thing by those being 
 displaced? Knocking on their doors might be considered an 
 invasion of their privacy. 
 
 Student: Maybe. But, data acquired by knocking on the doors of 
 those displaced people would allow, over some period of time, to 
 make more sophisticated and valid appraisals of whether to 
 redevelop an area, and what tertiary problems to expect when 
 these people are displaced. To me, that is a legitimate social 
 concern. This gets back to the point about the compromise 
 between the individual's privacy and the value of the potential 
 social indicator. In this instance, I think displacement is very 
 definitely a social indicator that we are going to have to pay more 
 attention to. 
 
 Student: In effect, you're coming to their door and saying, "We 
 
 60 
 
 want to help you, and we want to learn how we can help you." 
 
 Mr. Barabba: I wanted to make this point that once in a while 
 you have to make a tough decision as to whether you're willing to 
 give up a little piece of your privacy. 
 
 These are the kinds of things you've got to address yourself 
 to when you start looking at the information as a planner, or 
 whatever you might be. What's the cost of getting that informa- 
 tion? It's the cost of the information itself, and the cost to 
 society of having somebody knock on its doors. 
 
 Mr. Hanel: Do you remember in Midtown there was an area 
 where formal vacancies were increasing at the same time doubling 
 up was increasing, and the net result was more households rather 
 than less? The whole thing was unplanned, and there was a big 
 shift in demography. 
 
 It's one thing to have a planned development. It's another 
 to have a situation like Midtown's which just sort of happened, 
 and if you recall, it happened very suddenly. There was no 
 significant physical deterioration in that neighborhood, to speak 
 of; it was far more perceptual than physical. Different types of 
 people were moving in and they're coming in from somewhere, 
 yet, nobody quite knows where. Unless the people in Midtown 
 can deal with that, unless they can understand it, unless they can 
 encompass it in their planning and their thinking, that could 
 easily become a very difficult situation in that part of town. 
 
 What I submit is that in this kind of a situation, if you 
 don't understand these things, how in the world are you going to 
 deal with them? The compromise may very well be that you have 
 to know certain things, and the decisionmakers can profit from 
 hearing what the people have to say, if you can get them to talk. 
 
 Student: Could you comment on the changes of housing types 
 that we've had in the last 5 years? The move toward mobile 
 homes, condominiums, etc. How has that affected the breakdown 
 of housing? 
 
 Mr. Hanel: There is continuing and increasing pressure on 
 low-cost housing. While your housing stock may not be changing, 
 the number of households, the people who are competing for 
 available housing, is increasing. You have the pressure toward 
 more household units on one side and a lid on availability, mainly 
 because of money and cost, on the other. I think there's going to 
 be one dickens of a squeeze on central city housing stock. The 
 flight to the suburbs, in the light of the energy situation, is not 
 the long-range answer. 
 
 Student: Does the size of the city affect the reliability of what 
 you can do? 
 
 Mr. Hanel: No. The biggest one we've done is Houston and that's 
 1.25 million people. The smallest is around 12,000 people. Size 
 hasn't seemed to influence the reliability of the data. 
 
 Student: Why haven't you done Los Angeles? 
 
 Mr. Hanel: You can't make ends meet publishing a city directory 
 in Los Angeles. 
 
 Mr. Barabba: We pointed out that the production cost of this 
 type of data may remain stable, or relatively stable, overtime. It 
 
PROFILES OF SMALL URBAN AREA CHANGE 
 
 could go up a little bit or it could go down a little bit. One of the 
 things that keeps this price down is its use for other purposes. In 
 the case of the R. L. Polk Company, they sell the lists. 
 
 Mr. Hanel: An important correction regarding privacy; we don't 
 sell lists, and certainly we don't sell the records from which 
 particular lists are selected. We list from the files those addresses 
 at which we find the conditions or characteristics of interest to a 
 particular marketer, and we use that list for him under a one-time 
 rental arrangement. In most cases, the list never leaves our 
 premises. 
 
 We never search the files for the names and addresses of 
 individuals except when the individual has requested his record be 
 deleted from the file. 
 
 In any event, with only one exception, we never let the 
 basic information out of our hands. That exception is to supply 
 some city governments information as backup to the kinds of 
 statistics we're talking about today. Some cities want to do 
 subsampling, for example, among new movers or vacant housing 
 units. Or they want to put the numbers together for different 
 kinds of geography than tracts— for the length of a street, for 
 example. When we make address-by-address detail available, it's 
 under an agreement that the city will strictly control the use of 
 the file under its normal confidentiality restrictions. 
 
 Ironically, everything in our files is either from a public 
 record or represents the objective, nonsensitive, public face of 
 the individual. The real question in this whole business of lists 
 and directories and similar data is "What's private and what's 
 public?" It seems to me a gross extension of privacy to cover 
 every aspect of a person's identity, including those that can 
 readily be observed. As Mr. Barabba has said, being a member of 
 society implies being counted, and it also requires being identified 
 for various public and business purposes. 
 
 Mr. Barabba: You can also buy a directory of the people of that 
 city, published in a book, and that has what information in it? 
 
 Mr. Hanel: Everything we've talked about, except the number of 
 children. We don't publish that. 
 
 Student: The Census Bureau represents a public agency that is 
 policed by the Congress. R. L. Polk is a private agency, and there 
 doesn't seem to be anyone directly policing your standards of 
 confidentiality. 
 
 Mr. Hanel: It's important to bring out the fact that we publish 
 around 1,200 directories which include 7,000 communities, and 
 you can go back in time. Some of these have been published for 
 20, 30, 50, 70, and even 80 years. Also, there's nothing in those 
 books today that wasn't in them 10, 20, or 30 years ago. Not 
 only is the information the same, the procedure, with one 
 exception, has not changed either. However, it's that exception, 
 the use of computer processing, that needlessly scares everyone. 
 The computer doesn't make any difference in the content of the 
 directory. 
 
 Besides, I don't see any information we gather that could 
 be misused. As I said, we publish all the information we collect, 
 except the number of children in the family. 
 
 We go door-to-door with a carefully controlled survey. We 
 go, into each city each year almost on the same schedule. In a 
 city of 100,000 people it might take us 2 or 3 months to make 
 the canvass. The supervisory crew moves in, and they usually are 
 
 the same people who have been dealing with the city for years. 
 The interviewers are local people, usually neighborhood people- 
 housewives, retired people, and students in the summertime. 
 
 The canvassers are armed with a printout of the informa- 
 tion which appeared in the previous year's directory. It shows 
 street name and the list of all the addresses on that street, such as, 
 "105 Main Street is Smith, Jim. Wife's name Mary, etc." Their 
 job is to verify, update, and correct as necessary the record which 
 appeared in the previous year's directory. Actually, there is a 
 change of some kind in about 60 to 80 percent of the records. 
 They bring those work sheets in and we keypunch the changes. 
 Those corrections are fed into the computer, and that becomes 
 the data base for printing the next directory. 
 
 People ask us how we make sure that the interviewers don't 
 spend their time in the bar, or the drug store, or home reading, or 
 something. First of all, the supervisors call some respondents and 
 ask if the interviewer was courteous. Also, the computer is 
 programmed intentionally to insert a bit of nonexistent informa- 
 tion in about 1 in every 20 records, randomly, like the name of 
 someone who isn't there. The interviewer's work is checked to see 
 whether they have verified any of these false records. It's very 
 effective self-policing. 
 
 Mind you, one of the most important ways that the 
 accuracy of the information is policed is the fact that it is 
 published. It is there for people to use, read, and look up. Our 
 directories couldn't have lasted all these years if they hadn't 
 stood up under this kind of scrutiny. 
 
 Student: One of the hardest things for me to accept is the fact 
 that you're listing me in the directory. 
 
 Mr. Hanel: You might be interested to know we have not seen, 
 over the last several years, any appreciable change in the number 
 of people who ask not to be listed in the directory. You may note 
 in one of the tables in our reports, we show the number of 
 occupied households, the number from which we get interviews, 
 and the number where we get name only, or that ask us not to 
 show anything. And that runs only about 2 percent of the 
 total— a very small amount. 
 
 Student: Are the people given the alternative not to be listed? 
 
 Mr. Hanel: Oh, sure. 
 
 The question that you raise keeps coming up: "Do we 
 adequately inform the people of what we're doing?" Our answer 
 has stood up well, and that is, we do. That doesn't mean they 
 listen every time, and certainly we can't take out 5 minutes to 
 explain all the possible uses of the information, but there's no 
 attempt and no suggestion of any attempt to hide or in any sense 
 disguise the purpose of this solicitation. 
 
 Student: Have you ever had legal hassles? 
 
 Mr. Hanel: I remember a case about 3 weeks ago where a doctor 
 had asked not to be listed and somehow he was. He wrote us a 
 letter threatening to sue. If someone asks not to be listed in the 
 book, we comply. If there is any of that kind of problem, like the 
 doctor's, it's because of a slip-up that occasionally happens. But, 
 it's certainly not a matter of intent or neglect. One year the 
 whole police force of a certain city asked to be left out of the 
 directory, and we complied. 
 
 61 
 
PROFILES OF SMALL URBAN AREA CHANGE 
 
 Student: I, myself, have never been contacted by anyone for a 
 city directory, yet I have found myself listed in city directories 
 and I wonder where they got the information. 
 
 Mr. Hanel: Sometimes we get name lists from employers in the 
 case of public services. 
 
 Student: You see, I was never given the opportunity to refuse, 
 and I resent that. I wouldn't mind giving the information, really, 
 from the statistical standpoint. I just wouldn't want my name to 
 be published in any book. 
 
 Mr. Barabba: Will you accept the information and not publish it? 
 Is there an option that way? 
 
 Mr. Hanel: If we had the record it would be there. If you asked 
 not to be in the directory, we would obliterate your record. 
 
 Mr. Barabba: But you wouldn't carry his data forward? 
 
 Mr. Hanel: No. He would become what we call a "no return" at 
 that street address. 
 
 Student: Would it be necessary to interview him the next time? 
 
 Mr. Hanel: We wouldn't know that we were coming to him. We 
 would have his address with a "no return. "They'd knock on that 
 door again and he would say, "Well, last year I asked not to be 
 listed." The interviewer would know immediately at that point. 
 Now, if someone had since moved in at that place, they might 
 volunteer the information. 
 
 I think most of you would agree that year-to-year infor- 
 mation is an important dimension in understanding what is 
 going on in our cities. If it is, there's only one way to get it. And 
 it is not through the Bureau of the Census because the budget 
 people have said that the Nation can't afford a mid-decade 
 census, much less an annual census. The kinds of statistics that 
 we're talking about in understanding the dynamics of a city can 
 only come as a low-cost byproduct of an ongoing enterprise. 
 
 We started this venture about 1968 because we were aware 
 we possessed what amounted to an annual census, although we 
 were simply tabulating inventories. In 1971 we started working in 
 the dimension of the detailed components of change. 
 
 There are many things that such information might be used 
 for. If you want the answers to problems associated with the 
 energy crisis, employment, business opportunities, or the whole 
 city-suburban problem, then determine the sources of employ- 
 ment and residence. 
 
 Until people running cities learn how to understand and 
 deal with year-to-year small area change, they'll be reacting to 
 existing problems rather than anticipating and managing prob- 
 lems before they arise. 
 
 You talk about trade-offs. If you consider that this is valu- 
 able information, you must accept that there is no way to obtain 
 it at these costs, except to do precisely what we're doing. 
 
 Mr. Barabba: Let's assume a congressman introduces a bill that 
 requires any organization with individuals in a computer file to 
 apprise those individuals that they are in the file, what is on the 
 file, and that they be notified everytime that file is used. So you 
 have a hearing and everybody tells you about how good it is to 
 protect the individual. Mr. Hanel comes in and says, "Well, 
 
 instead of now being able to give you all this information for 15 
 cents, it's going to cost you $1.50." It's also going to cost 10 
 times more to get a mailing list to communicate for marketing 
 purposes, it's going to cost 10 times more to get at this kind of 
 information relative to community analysis. 
 
 Student: Is such a bill actually pending in Congress? 
 
 Mr. Barabba: Yes. 1 
 
 Student: Originally we talked about using the information for a 
 purpose that was not intended when the information was given. 
 Our information here is going to be published. If somebody buys 
 the book and uses it for an entirely different purpose, that's O.K. 
 Anybody can buy the book, and people knew they were giving 
 information to be published. 
 
 I think that it might be appropriate, if I allow my name to 
 be used, that Mr. Hanel have a signed release from me, and on the 
 release it would explain where my information was going. 
 
 Mr. Hanel: You've just blown us out of the water cost-wise 
 because you start to do that and we are into a 5-minute 
 conversation. 
 
 Student: Probably, but you indicated that there may be as much 
 as a 25-percent turnover in some cities. So, that release could be a 
 continuing thing until a letter was written to you to delete that 
 information. 
 
 Mr. Hanel: That means filing and handling 27 million letters. Not 
 to mention the cost of originally getting the letters or signed 
 forms. We once asked a question on the ages of children as an 
 experiment when some city wanted that. People are used to 
 furnishing certain information for the city directory, and they 
 say, "Why are you asking that for the first time this year?" It 
 doubles or triples the interview time and cost right there. 
 
 Student: I agree that it has a price, but most people every year 
 have to file an income tax return. It would seem to me that 
 people could also file some type of statistical data every year, 
 that wouldn't be used by name, so they could come up with 
 different information. 
 
 Mr. Barabba: We call that a census, only we do it every 10 years. 
 
 Student: But the income tax thing is all done by mail. The 
 people do it, and why couldn't they do the other? 
 
 Mr. Barabba: Because, it would cost over $300 million per year 
 to receive, process, tabulate, and evaluate all the citizens. 
 
 Well, let's assume you are congresspersons, what do you 
 think? Do we ask Mr. Hanel to change the way he does things? 
 
 Student: I don't think it's an invasion of privacy just to make a 
 list of people's names. If that list were mandatory then it might 
 be an invasion of privacy. Otherwise, it's strictly a voluntary 
 thing, I can't see where it's any invasion of privacy and I'm sure 
 if anybody wants to find out who lives in such-and-such a 
 household, he can find out anyhow. So you might as well make it 
 easy for him and give him all the statistics. 
 
 Privacy legislation regulating Federal activities was enacted December 
 31, 1974-Public Law 93-579, Section 3418. 
 
 62 
 
PROFILES OF SMALL URBAN AREA CHANGE 
 
 MAP 1. MILWAUKEE, WISCONSIN 
 
 PERCENT OF OCCUPIED HOUSING UNITS 
 
 WITH CHANGE OF FAMILY 
 
 27.35 
 
 50% OR MORE 
 
 25% OR LESS 
 
 63 
 
PROFILES OF SMALL URBAN AREA CHANGE 
 
 W 
 
 H 
 
 Z 
 D 
 
 z z d 
 o 55 5 
 
 h o 2 
 <i u. 
 
 I Q ° ca 
 
 £ £ ui 8 
 
 O Q. (5 co 
 Z D Z " 
 
 O " < 
 
 OX 
 
 O U 
 esi u. T 
 
 0. O H 
 
 i^i 
 
 UJ 
 
 O 
 
 DC 
 
 UJ 
 
 0. 
 
 I 
 
 64 
 
PROFILES OF SMALL URBAN AREA CHANGE 
 
 MAP 3. SAN ANTONIO, TEXAS 
 
 PERCENT OF OCCUPIED HOUSING UNITS 
 
 WITH CHANGE OF FAMILY 
 
 36.78 
 
 50% OR OVER 
 
 40% - 49% 
 
 25% OR LESS 
 
 65 
 
66 
 
PROFILES OF SMALL URBAN AREA CHANGE 
 
 MAP 5. PERCENT OF OCCUPIED HOUSING UNITS 
 
 WITH CHANGE OF FAMILY 
 
 27.70 
 
 50% OR MORE 
 40% - 49% 
 25% OR LESS 
 
 67 
 
PROFILES OF SMALL URBAN AREA CHANGE 
 
 MAP 6. NET CHANGE IN TOTAL HOUSING UNITS 
 
 1,216 
 
 68 
 
 25-49 
 50-99 
 
 100 OR MORE 
 
PROFILES OF SMALL URBAN AREA CHANGE 
 
 MAP 7. NET CHANGE IN VACANCIES 
 CHANGE IN OCCUPANCY OF EXISTING UNITS 
 
 + 437 
 
 3000 6000 F 
 
 INCREASES 
 
 DECREASES 
 
 TRACTS WITH INCREASE 
 OF 25 OR MORE HOUSING 
 UNITS - OTHER THAN NEW 
 CONSTRUCTION 
 
 69 
 
PROFILES OF SMALL URBAN AREA CHANGE 
 
 MAP 8. NET CHANGE IN HOUSEHOLDS 
 + 788 
 
 _i i 1 
 
 25-49 
 50-99 
 100 OR MORE 
 
 70 
 
PROFILES OF SMALL URBAN AREA CHANGE 
 
 MAP 9. NET CHANGE IN JOBLESS HEADS OF HOUSEHOLD 
 (EXCLUDING RETIRED & STUDENTS) 
 + 702 
 
 3000 6000 FEET 
 
 25-49 
 
 71 
 
PROFILES OF SMALL URBAN AREA CHANGE 
 
 MAP 10. PERCENT HOMEOWNERS 
 66.62 
 
 70% - 89% 
 90% OR MORE 
 
 72 
 
PROFILES OF SMALL URBAN AREA CHANGE 
 
 MAP 11. HOUSEHOLD INCOME 
 
 AVERAGE INDEX POINTS 
 
 102 
 
 112 OR MORE 
 I 92 OR LESS 
 
 73 
 
PROFILES OF SMALL URBAN AREA CHANGE 
 
 MAP 12. NUMBER OF CURRENT-YEAR VACANCIES FOR SALE 
 OVER/(UNDER) MINIMUM REQUIREMENTS 
 
 299 
 
 15 OR MORE UNITS 
 OVER MINIMUM 
 
 74 
 
 LESS THAN MINIMUM 
 LOW-INCOME AREAS 
 
PROFILES OF SMALL URBAN AREA CHANGE 
 
 MAP 13. NUMBER OF CURRENT-YEAR VACANCIES FOR RENT 
 OVER (UNDER) MINIMUM REQUIREMENTS 
 
 809 
 
 i i i i 1 1 
 
 15 OR MORE UNITS 
 OVER MINIMUM 
 
 LESS THAN MINIMUM 
 
 LOW-INCOME 
 '/ AREAS 
 
 75 
 
PROFILES OF SMALL URBAN AREA CHANGE 
 
 MAP 14. PERCENT OF HOUSING UNITS VACANT 
 
 6.86 
 
 10% -19% 
 
 20% OR MORE 
 
 Yj 
 
 LOW-INCOME AREAS 
 
 76 
 
THE ROLE OF PRIVATE FIRMS IN DEVELOPING INDICATORS 
 
 Richard Wirthlin, President, Decision Making Information, Santa Ana, California 
 
 "If we had simply taken the aggregate data we would have been more than half 
 blind. When you begin separating the data by city, only then does it have significance in 
 terms of a public administrative action that makes sense for residents of Compton, 
 Paramount, Lynwood, Bellflower, and Downey." 
 
 Richard Wirthlin 
 
 Mr. Barabba: I'd like to introduce Dr. Richard Wirthlin of 
 Decision Making Information, a national survey research firm. 
 Dr. Wirthlin's academic background is in the area of economics, 
 and he has been chairman of that department at Brigham Young 
 University. Prior to helping form Decision Making Information 
 (DMI), he headed another firm, Merrill/Wirthlin and Associates, 
 which was strictly a survey research operation. DMI was formed 
 when Merrill/Wirthlin merged with Datamatics, Inc., an organi- 
 zation basically involved in the computerization and analysis of 
 secondary data. Dr. Wirthlin intends to discuss the type of 
 assistance that such a private firm can provide decisionmakers as 
 well as the techniques used to accomplish the survey research. 
 
 Mr. Wirthlin: It's a great pleasure to be with you and to share a 
 session with my good friend, Mr. Barabba. 
 
 The general topic of this entire series is, I think, extremely 
 timely— developing social indicators to cope with "future shock." 
 Kenneth Boulding, an economist, has said, "We live in a firestorm 
 of change." This day, August 9th, has brought a dramatic 
 example of how rapidly political change can occur. A new 
 president assumes office, it seems, almost overnight. 
 
 Someone, I think it was the German philosopher Goethe, 
 once said, "What is not understood is not controlled." We can 
 paraphrase and say, "What is not understood cannot be admin- 
 istered well." If there is an ultimate social objective to gathering 
 and processing either primary or secondary data for public 
 administrators, it is, I believe, to understand the forces that shape 
 and mold society, so that we can better administer and serve 
 those people affected by societal changes. 
 
 I would like to address myself to what a private research 
 firm can do for you by detailing some of our activities at Decision 
 Making Information. 
 
 But, first, let's briefly identify the major advantages of 
 working with survey research data (as opposed to secondary 
 data), discuss its logic and analysis, and demonstrate how it 
 relates to, and can strengthen, secondary data. 
 
 The analytic strengths of survey research data are first, 
 inasmuch as you need not interview all individuals, the data are 
 timely— a survey can generate data within days and, in some cases, 
 within hours. Second, topically, survey data can be as focused or 
 as diffused as desired. One might say it provides the public 
 administrator with a zoom lens. He can pull back and view an 
 entire city on an aggregate level, or he can zero in on a particular 
 subgroup within the community, sampling from it to determine 
 the specific and analytically critical characteristics of that 
 subgroup. 
 
 For instance, the number of female heads of households 
 who have moved within the last year could constitute a universe \, 
 
 which is simply a collection of individuals who have a specific 
 characteristic. From that universe we select a random sample 
 which generates the information of interest. The basis of all 
 survey research is proper sampling of a carefully identified 
 universe. 
 
 Third, survey research provides us with an understanding of 
 the nature and strength of relationships not obvious or even 
 discernible when we use aggregative or cumulative data. This 
 afternoon I'd like to overview the nature and application of 
 survey research. 
 
 About the assigned readings, the Office of Management and 
 Budget book, Social Indicators, 1973 edited by Raymond A 
 Bauer, should provide some historical perspective of social 
 indicators as generated and evaluated by survey sampling. The 
 analytic mode of survey research data is reviewed in The Logic of 
 Survey Analysis by Morris Rosenberg.' 
 
 Let's now review the logic and implementation of survey 
 analysis. As I said earlier, one great advantage of doing primary 
 research lies in the fact that you can select the specific building 
 blocks of the analysis. You are not restricted to looking at only, 
 say, income, occupation, or vacancy rates. You can develop a 
 wide array of variables and use them to test a given hypothesis 
 about society. Even more important, survey research enables the 
 researcher to go beyond the standard descriptive variables. It is 
 critical for classification purposes to deal with demographic 
 variables such as age, number of children, education, etc. If we 
 designate age as the first demographic variable of interest D| , and 
 the number of children as D 2 , then there is a large but finite 
 number of demographic variables, the last of which we'll call Dn. 
 Thus we may select for analytic purposes any of the Di, 
 D 2 , . . . Dn variables. However, in most survey research packages 
 we also deal with certain economic variables such as occupation 
 and income. Again there would be an "n" number of these 
 economic variables that we would want to consider; these can be 
 designated as E , , E 2 , . . . En. 
 
 We may deal with certain political variables such as political 
 alienation, efficacy, apathy, cynicism, the attitude the respondent 
 may have about a given bond issue, or his propensity to vote. 
 These variables may be identified as P, , P 2 ,. • • Pn, Thus, there's 
 an "n" dimension of political variables we can generate and use in 
 our survey analysis. Possible territorial or ecological variables T 1( 
 T 2/ . • ■ Tn may be of interest. There are also more general 
 attitude variables that we may want to crank into our initial 
 hypotheses such as attitude toward work, status, self, etc. (A 1( 
 A 2 , . . . An). In other words, we can deal with an almost infinite 
 number of dimensions for the purpose of analysis when we begin 
 
 'Morris Rosenberg, The Logic of Survey Analysis, Basic Books, Inc. 
 New York, 1968. 
 
 77 
 
THE ROLE OF PRIVATE FIRMS IN DEVELOPING INDICATORS 
 
 constructing the questionnaire or, as we call it, interview 
 schedule. 
 
 There are some constraints once you determine the 
 hypothesis. The first and most fixed constraint on the number of 
 information pieces you can gather in any one interview is the 
 amount of time available for the interview itself. Generally, we 
 can interview a respondent for about an hour at any one sitting. 
 Sometimes the interview can last as long as an hour and a half. 
 But, as a rule of thumb, we find that response error and the 
 number of nonresponses are kept pretty well in line if we keep 
 the interview to 60 minutes or less. 
 
 Student: Our instructor has shown us a graph apparently 
 indicating that, if you can get a couple of variables that will give 
 you more information, you're just as well off when you're using a 
 few rather than using a large number in the analysis. Do you find 
 that with more variables you can do a better job with the analysis 
 or does it hold that with the fewer variables and a more brief 
 interview you can come up with the same answer? 
 
 Mr. Wirthlin: Clearly, the study will cost less if the survey 
 generates only variables that statistically are highly explanatory. 
 The major limitations in identifying these particular variables lie 
 in the fact that frequently, unless we simply replicate someone 
 else's study, we don't know beforehand what these explanatory 
 variables might be. 
 
 Student: Do you add more variables just in case somebody 
 challenges you later on, so you can be defensive and say, "We 
 have examined these other parameters?" 
 
 Mr. Wirthlin: I prefer to view the value of having more variables 
 available for analysis not in terms of being defensive, but rather 
 having a margin of analytic safety. The strength and flexibility of 
 survey research lie in the wide range of possible variables. We can 
 create the variables that we think, a priori, are important to the 
 problem with which we're dealing, be it crime, health statistics, 
 mobility, or identifying an agenda for change in a community. 
 
 Student: Do you prefer, from your own experience, a few or a 
 lot of questions? 
 
 Mr. Wirthlin: If you and I lived in a world of unlimited resources, 
 then clearly I would prefer "a lot." But we don't. Further, the 
 number of variables used in the analysis depends on the 
 complexity of the task. In some cases, we can get all the 
 information we need and more in a matter of 5 minutes. In other 
 cases, it might take us several hour and a half sessions to 
 ferret out the variables of potential importance. Toward the end 
 of our session today I'm going to talk about a study that we did. 
 Measuring Outcomes of Manpower Programs. Virtually every one 
 of the variable dimensions that I've listed, demographic, eco- 
 nomic, political and attitudinal, were used in that study because 
 "outcomes" could not be measured simplistically and dimen- 
 sionally in terms of the training experience alone. Individuals who 
 went through training had vastly different experiences both 
 before and after their training. Now, you have to control— and 
 we'll get into this when we talk about the analysis of survey 
 research— a broad series of test variables. Age determined, for 
 instance, the potential employability of an individual. Education 
 surely did. Attitude variables also influenced outcomes. There- 
 
 fore, we had to look at a very broad spectrum of possible 
 employment-determining elements that might affect an individ- 
 ual's work experience. 
 
 Student: When you plan the interview, do you have some 
 distractive variables to put the respondent at ease so you get 
 better information from him? 
 
 Mr. Wirthlin: Generally, we initially ask something that invites a 
 quick and painless response. One interview we're using right now 
 in Orange County is designed to identify travel patterns to assist 
 the county transit district in setting up its vehicle scheme 
 efficiently. That interview begins this way: 
 
 Hello, I'm so-and-so of Decision Making Information, a 
 national research organization. We're conducting a survey in 
 your neighborhood today concerning public transit in 
 Orange County. Do you have a current driver's license? 
 
 And with that, we're right into the interview. If the person 
 responds to that first question, almost always he will complete 
 the interview. 
 
 There's a great deal of art in this so-called science of survey 
 research— properly pacing the interview, using visual aids to keep 
 attention and interest high, employing a variety of questioning 
 techniques. These are all methods employed to get high response 
 rates. 
 
 The other constraining factor on our "unlimited" array of 
 variables lies in the accuracy of responses to the questions. There 
 are certain response errors which clearly limit the scope of the 
 data that can be acquired through the survey method. For 
 instance, if you asked a person whether he was registered to vote 
 in the 1972 election, and, if so, did he vote, you'll find that most 
 people, because they feel it's a social imperative kind of thing, 
 will say they voted, when, in point of fact, they didn't. For 
 instance, in national post-election studies conducted in November 
 1972, about 82 percent of those polled said they had voted, while 
 the national statistic is actually around 60 percent for voter 
 turnout. So you must be very much aware of the limitations 
 inherently imposed by this method of data collection when you 
 initially establish your array of analytic variables. 
 
 The Logic of Analysis 
 
 Now, what is the logic of survey analysis? We have a limited 
 number of possible factors. Once the data is gathered, we then try 
 to identify those variables that are associated with each other. 
 We can't identify causations, but, by introducing a test variable 
 such as age or attitude, we can ferret out the strength of the 
 particular relationships we are interested in between the training 
 experience and change in income, exclusive of age or attitude. 
 
 Hence, the entire analysis and, thus, the structure and scope 
 of the questionnaire are conditioned by your hypothesis about 
 the relationship between two or more variables. On this base, you 
 develop your interview schedule or questionnaire. Some questions 
 are closed-ended, responses for which are limited and prede- 
 termined. For example, "Do you have a drivers license?" (yes 
 or no) or "Will you vote for candidate A or B in the election for 
 Senator?" (A, B, or undecided). Other questions are open-ended, 
 or have no preset responses and no limits. For example, "What do 
 you consider to be the most important problem facing your city 
 today?" 
 
 78 
 
THE ROLE OF PRIVATE FIRMS IN DEVELOPING INDICATORS 
 
 One type of open-ended question we have found to be of 
 great help in understanding the values of a community is based on 
 Bavelas' "echo" technique: "What would be a good thing to 
 happen here in the city of Los Angeles?" followed by "Who or 
 what could cause it to happen?" This very simple questioning 
 format generates an elegant variety of values related to the object, 
 in this case the city of Los Angeles. One ancillary payoff of this 
 particular technique is capturing the verbatim responses, and, 
 thus, the tone and semantic nuances people use to describe their 
 feelings about a given object. 
 
 We used this questioning approach almost exclusively in a 
 study of a San Jose, Calif., ghetto. The city government initiated 
 some community action programs which failed miserably. It felt 
 it could not get honest and clear feedback using the standard 
 channels of forums and committees. So we surveyed, using echo, 
 and got a wide variety of opinion as to what people wanted to 
 happen in their area, and who they felt could make it happen, as 
 well as the flavor of their expressed concerns about their 
 
 community. 
 
 Another interesting technique we use is something we call 
 "Alias Smith and Jones," which goes like this: 
 
 Mr. Smith believes that schools are inadequately financed 
 and, therefore, favors a bond issue. Mr. Jones believes that 
 taxes are already too high and opposes any school bond 
 issue. Other things being equal, do you consider yourself to 
 be more like Mr. Smith or Mr. Jones? 
 
 Using this format we can elicit an opinion on any highly 
 controversial issue without biasing the respondent's answer. From 
 this particular question we could develop a pro/con education 
 scale. 
 
 A variety of other scaling techniques may be used. Another 
 one we're using in Orange County goes like this: 
 
 I'd like to ask you some questions about various bus rates 
 and time schedules that might be offered. For each 
 situation I read, tell me how many times out of 10 trips 
 you might ride the bus to (destination) for 1) work, 2) 
 school, and 3) shopping. Here's the first situation. If there 
 were a bus system that was free and ran one block from 
 your home and reached its destination every 15 minutes, 
 how many times out of 10 trips would you ride it? 
 
 They respond to this situation for work, school and shopping, 
 and then we continue through a series of situations, such as: 
 
 If it were 35 cents, ran one block from your home, reached 
 its destination every 30 minutes. . . . 
 
 Thus, by altering price and time through the series, we can derive 
 insight into how people value convenience versus economy. 
 
 Student: You talk about measuring an attitude toward riding the 
 bus, but who's to say someone's expressed attitude equates to 
 putting bus fare in the box? 
 
 Mr. Wirthlin: Good question. In fact, a person may not ride the 
 bus, despite the favorable response to a single question. That's 
 one reason why this is one of those interviews that takes an hour 
 and a half to conduct. 
 
 Remember my example of asking whether someone voted 
 in the November election? You can't always accept people at the 
 face value of their responses. But we can still determine, with 
 relatively high accuracy, whether he really did vote by asking him 
 other questions— about past behavior, interest in political cam- 
 paigns, and knowledge of the election process and the candidates. 
 From these responses we can develop a vote probability scale. 
 
 The same is true for the potential bus rider— we ask another 
 series of questions to delineate the gap between what he says his 
 attitudes toward rapid transit are and what, in fact, they are. 
 
 The Likert technique is another questioning device that we 
 find very fruitful, particularly in measuring controversial atti- 
 tudes. The Likert format runs counter to the basic training of a 
 survey researcher which emphasizes that good questions must be 
 scrupulously unbiased in their influence upon responses. Rather 
 than asking the individual a neutral, balanced question, we read 
 him a blatantly opinionated or provocative statement; then ask 
 him to indicate whether he agrees strongly, just agrees, disagrees, 
 or disagrees strongly with the expressed sentiment. Let me give 
 you some examples: 
 
 "People like me don't have any say about what local 
 governments do. Do you agree strongly, agree, disagree, or 
 disagree strongly with that statement?" 
 
 Another one: "It would be a big adjustment for me to ever 
 use public transportation." 
 
 Student: When you use the Likert items does the respondent see 
 the statements on a card? 
 
 Mr. Wirthlin: Yes. 
 
 Student: I could see if you're doing this orally there might be 
 grounds for some confusion. 
 
 Mr. Wirthlin: Exactly. The great advantage of interviewing 
 face-to-face is that you can employ visual displays. You can use 
 what we call card helps so that as the interviewer reads the 
 questions the respondent also sees it, and he can determine 
 whether he agrees strongly, agrees, disagrees, or disagrees 
 strongly. 
 
 Still another technique— and one which depends on this 
 visual assistance— is the Semantic Differential. The semantic 
 differential measures image or perception of an object using 
 dichotomous adjective pairs. For example, in measuring percep- 
 tions of bus travel characteristics, we show respondents a card 
 with adjectives such as: 
 
 Comfortable 1 2 3 4 5 6 7 Uncomfortable 
 
 Then we ask them to rate bus service on that scale. If they think 
 that traveling the buses is extremely uncomfortable, they would 
 mark a "7"; extremely comfortable, a "1"; or something in 
 between the two. The key here is for a respondent to go down 
 the list of dichotomous adjectives quickly because "top-of-mind" 
 responses give the most accurate index. In this study some of the 
 other adjective pairs were late/on time, simple to use/compli- 
 cated, safe/dangerous, old fashioned/modern, slow rush-hour 
 trip/fast rush-hour trip, high status/low status, convenient/incon- 
 venient, unenjoyable/enjoyable, expensive/inexpensive, unreli- 
 able/reliable, and relaxed/tense. Then we asked them to judge the 
 same travel characteristics for automobiles. 
 
 By comparing the mean or average responses to these two 
 scales, we can profile each of those adjective pairs and generate a 
 map, if you will, of the image attributes of bus travel versus car 
 
 79 
 
THE ROLE OF PRIVATE FIRMS IN DEVELOPING INDICATORS 
 
 travel. We might find that the automobile is viewed as a more 
 comfortable and faster vehicle, but the bus is viewed as less 
 expensive, and so on. Such image responses, in combination with 
 many others, can help define the perceived problems of using the 
 bus in Orange County. You can use this technique in developing 
 perceptions vis-a-vis health care delivery systems, the police force, 
 a political candidate, the community as a whole, or for any object 
 or person to develop some kind of image profile. 
 
 Student: Why is a 7-point scale necessarily so much more 
 sophisticated than a 5-point scale? I don't see that what you get 
 from 7 would be more helpful. 
 
 Mr. Wirthlin: A lot of research has gone into the psychology of 
 testing, and the number of points on a scale are a simple trade-off 
 between ease of use and precision of measurement. People can 
 handle a 7-point scale with relative ease when it's presented 
 visually. Of course a 5-point scale is somewhat easier and quicker, 
 but even that won't work over the telephone. In that case we are 
 limited to 3 points. 
 
 If you had a sample of 384 individuals your confidence 
 interval would be ±5 percent. At 600, it's 4 percent. At 1,536, it's 
 about 2.5 percent. Now look what happens. To buy, if you will, 
 one degree of accuracy, or to shave off 1 percent of our 
 confidence interval, we only had to interview 200 more individu- 
 als as we moved from ±5 to ±4 percent. To buy the next degree 
 of accuracy of ±1 percent we would have to interview an 
 additional 936 individuals. Now suppose we want an interval of 
 ±2 percent; we would have to interview 2,400 individuals or an 
 additional 864. If we wanted a confidence interval of only 1 
 percent, we have to sample 9,600 individuals— an additional 
 7,200. The relationship between required sample size (given the 
 95-percent confidence interval) and accuracy becomes almost 
 infinite below the ±1 percent level. In other words, the accuracy 
 of the sample given economic constraints for most surveys 
 dictates a sample size of about 1,500 to 2,000. We deal with 
 considerably larger samples, in some cases, to permit the 
 researcher to speak with a high degree of accuracy about certain 
 small subgroups in the defined universe, or when the need for 
 accuracy is demonstrably evident. 
 
 Sample Selection 
 
 Another critical step in the implementation of a survey, 
 concurrent with questionnaire development, is sample selection. 
 
 Without going into a great detail, let me say that the object 
 of good sample design is simple. Once you define your universe, 
 be it female heads of households who have moved within the last 
 3 months, or people displaced by a development, or victims of 
 a crime, or those who have had major surgery in a hospital, the 
 sample technique must do only one thing— give every individual in 
 that universe an equal probability or chance of being chosen. If 
 you do that consistently, you develop what's called a probability 
 sample. Slonim, in the assigned reading, goes through, in a rather 
 delightful fashion, a wide variety of approaches to developing and 
 implementing a random sample. The Census Bureau, by the way, 
 is foremost in developing stratified samples which are even more 
 precise than simple random samples. But you can boil all those 
 techniques down to one rule— everyone in that universe has the 
 same chance of being chosen. 
 
 Sampling is sometimes called the ABC science— the science 
 of Almost Being Certain. You're never 100-percent certain that 
 results from your sample will correspond with the results of 
 interviewing everyone in the universe. But the great advantage of 
 random sampling over other methods of sampling is that you can 
 precisely calculate the odds of a "mistake," and we call that 
 calculation a confidence interval. You can deal with any 
 comfortable confidence interval— a bigger one if you're a high 
 roller, and a smaller one if you don't like taking risks. Generally, 
 we like to deal with odds that are 95 out of 100. In repeated 
 sampling, there exists a different standard error for each sample. 
 Taken to its extreme, if we were to draw 100 samples and 
 calculate the 95-percent confidence interval for each sample, in 
 95 percent of our samples the interval will include the true or 
 universal mean, and 5 percent will not. That is to say, we are 
 95-percent confident our estimate of confidence interval does 
 include the true mean. Sometimes we want odds as good as 99 
 out of 100. After picking your odds, the width of the confidence 
 interval directly depends on sample size— the larger the sample, 
 smaller the interval. But does this mean we should take samples 
 of 5, 10, or 15 thousand? The answer is generally "no." Let me 
 show you why. 
 
 Mr. Barabba: Can you think of an example where you would 
 want to go below that ±1 percent? How about the unemployment 
 rate? You might want one-tenth of 1 percent. Right? The Census 
 Bureau interviewed 50,000 households to determine the unem- 
 ployment rate. Now you know the cost for that level of 
 confidence. 
 
 Mr. Wirthlin: Yes, you really pay for those last tenths of a 
 percent. 
 
 The second thing to remember about sampling is the size of 
 the universe really doesn't much influence the size of the sample. 
 If our universe or population is 10,000, we would have to inter- 
 view 964 individuals to get a ±3 percent accuracy. By interview- 
 ing only about 100 more, we can get the same accuracy for a 
 universe of 200 million, or, for that matter, half a billion. So the 
 percent that the sample represents in the universe becomes vir- 
 tually meaningless after the universe exceeds 10,000. 
 
 Student: Do most polls, like the Harris poll, interview about 
 2,000 people for a national sample? 
 
 Mr. Wirthlin: Between 1,500 and 2,000. And now you know 
 why. To buy 1 percent more accuracy, Harris would have to 
 increase his cost considerably. Therefore, unless you have a very 
 specialized problem and are in a position to purchase costly 
 survey research, be wary of very large samples. In most cases, a 
 sample of 1 ,500 will more than satisfy your needs, contingent, of 
 course, on well-designed random sampling techniques. 
 
 Student: Suppose you're interested in a universe of about 10,000 
 people, and suppose you have to stratify it, just for the sake of 
 discussion, into several separate groups. What rules of thumb now 
 hold in setting the size of the sample large enough to detect 
 whether there are differences between these groups? 
 
 Mr. Wirthlin: If you want to make statements about each 
 subgroup, then the requisite sample size depends upon how 
 homogeneous the population is. For example, a chef can sample a 
 huge bowl of soup with only one teaspoonful, but before he dips 
 his spoon he always stirs it so that the mixture is homogeneous. 
 
 80 
 
THE ROLE OF PRIVATE FIRMS IN DEVELOPING INDICATORS 
 
 Student: Let's take the case where the total is not very 
 homogeneous but each of the strata is homogeneous. 
 
 Mr. Wirthlin: Then you can get away with relatively few 
 interviews in each stratum. There are statistical techniques such as 
 chi square that can indicate if the differences between subsamples 
 could have occurred by pure chance and the mathematical 
 likelihood of being wrong. 
 
 Let's now proceed to the next steps of survey implementa- 
 tion. Briefly, the work of each interviewer is independently 
 validated. Then the interviews are coded to a series of numbers. 
 The computer checks the consistency of the respondent, and 
 generates a series of cross tables. These tables form the basis, a la 
 Rosenberg, for the analysis. 
 
 Primary and Secondary Data Linkages 
 
 Let's now discuss tne linkage between primary and second- 
 ary data and show how these linkages can be used to solve some 
 of the problems that you deal with. 
 
 A lot of the data that city planners work with is ecological; 
 it focuses on the study of territorial and spatial environments of 
 human activity and variations among units at some level above 
 the individual. Frequently, the use of secondary ecological data 
 makes it difficult to tease out the relationships between inde- 
 pendent and dependent variables. At the extreme, the association 
 of an ecological variable (a place variable, e.g. urban/rural) with 
 some other macro variable (political participation) can lead to 
 gross analytic error. 
 
 For example, if one examined aggregate ecological data 
 relating only urbanization and political participation in the 
 Western World over the last 175 years, a seemingly logical 
 inference is urbanization increased political participation. How- 
 ever, when survey research data including such items as educa- 
 tion, age, income, etc., are collected in these same areas, other 
 variables are found to influence voting behavior, and the logical 
 link is less fortuitous. 
 
 Let me give you a more dramatic example. In Denmark, 
 there is a close correlation between the number of babies born 
 and stork nests, per square mile. It would be ridiculous to 
 conclude one causes the other. But, we must ask why there is 
 such a relationship. The answer is population density. The houses 
 have chimneys, and where there are more chimneys, the proba- 
 bility of having a stork nest increases because storks nest in 
 chimneys. Chimneys also represent households. The more house- 
 holds, the higher number of babies born per square mile. So our 
 initial correlation is more a coincidence than a causal relationship. 
 The same thing happened in my first example. It wasn't 
 urbanization that caused increased political activity, but other 
 factors. 
 
 When you compare one global characteristic with another 
 global characteristic you expose yourself to the chance of 
 committing the ecological fallacy. By extrapolating from test 
 factors generated from individuals through survey research you 
 greatly reduce the risk of falling into this analytic trap. The 
 process of elaboration brings other variables into the analysis to 
 see whether, in fact, the relationship between the first two 
 versions holds or disappears. If it disappears, as it does when we 
 relate crime rates to racial characteristics, then we know that the 
 relationship is spurious. Thus, survey research can be used to 
 strengthen and illuminate whatever secondary data resources you 
 might have at hand. 
 
 In summary let's identify some analytic strengths of survey 
 research. First, it can identify your explanatory variables. That is, 
 it can say age and income, not race, are the major factors that 
 influence crime rates. 
 
 Second, it has the great advantage, as we discussed earlier, 
 of being able to focus on any particular subgroup. 
 
 Third, survey research can be used to measure the impact of 
 policy. The Orange County (Calif.) Transit study provides an 
 example. An initial study determines the attitudes toward transit 
 systems. A transit system is developed in accord with those 
 attitudes; we can then monitor the success of that program via 
 popular response. 
 
 Fourth, survey research can also be used to measure the 
 impact of situational change. We select a group of people and 
 measure their status, be it economic, political, or social. After 
 some change has been induced, this same group can be 
 reinterviewed, say, 3 or 4 years later, to determine the effect of 
 change. An example of this type of study is the evaluation of the 
 public employment program (PEP). Enrollees were interviewed 
 prior to their entry into the program and reinterviewed during it. 
 Now we're following up with a post-PEP interview to determine 
 where they are and what they're doing. 
 
 Fifth, you can also use survey research to measure the 
 acceptability of various alternative courses of action. Economists 
 have always avoided making value judgements about "socially 
 preferred" actions. Economic efficiency can be measured, but 
 how do you measure which of four economically efficient 
 policies is best for the public? Survey research can develop a 
 social value or welfare function that can help decide whether the 
 city fathers should build a park, widen the roads, or build new 
 schools based on need as defined by potential usage and desires. 
 
 Last, survey research can be used as a feedback mechanism 
 by continually monitoring the attitudes in a given community 
 about a wide variety of subjects. 
 
 Some Case Studies 
 
 Let me quickly go through some of the projects that have 
 been done in these areas and share with you some of the results. 
 
 In the spring of 1973 the California cities of Downey, 
 Bellflower, Lynwood, Paramount and Compton commissioned a 
 survey research project. One question asked was, "What do you 
 consider to be the most important problem facing your city 
 today?" If you look at the aggregate data (table 1), most were 
 concerned about crime and the police force; pollution and noise 
 rank a very poor second; teenagers and youth, third; and streets, 
 freeways and parking rank fourth. Measurable concern about 
 education, drugs, moral decay, shopping centers, racial problems, 
 pocketbook issues of taxes and inflation, and unemployment also 
 emerged. These categories are created from the free responses to 
 the question above. We didn't hand people a card and ask which 
 of these problems they felt were most important. We let them 
 respond in any way they wished, while the interviewer recorded 
 the verbatim responses. Later our coders wrote those responses 
 on cards and then placed the cards in logical categories— police, 
 crime, and so on— based on the analyst's subjective judgement. In 
 this case, the people said the crime rate was too high, therefore, 
 we need more police. The problem wasn't police brutality or 
 inefficiency. So, grouping them was analytically valuable, 
 although they could be separated in other studies. 
 
 Now, look at the differences that emerge when we compare 
 these concerns between the various cities. In Compton over half 
 
 81 
 
THE ROLE OF PRIVATE FIRMS IN DEVELOPING INDICATORS 
 
 of those interviewed identified crime as the most important 
 problem that city faced. Second was teenagers and youth. More 
 than 7 out of 10 people in that community were upset about 
 crime and youth-related problems. On the other hand. Paramount 
 residents were much more concerned about pollution and 
 education than crime. In Lynwood there were streets, freeways 
 and parking, and racial problems at the forefront. In Bellflower it 
 was shopping centers; in Downey, streets, freeways, drugs, and 
 moral decay. 
 
 Now if we had simply taken the aggregate data we would 
 have been more than half blind. When you begin separating the 
 data by city, only then does it have significance in terms of a 
 public administrative action agenda that makes sense for resi- 
 
 dents of Compton, Paramount, Lynwood, Bellflower, and 
 Downey. 
 
 Now let's look at the measured level of concern about the 
 drug and crime problems (table 2). In this case, we categorized 
 the data by age, sex, and geographic area. Clearly age differences 
 influence attitudes about drugs. Of those 18 to 24 years old, 55 
 percent said they were very concerned about drugs, in contrast to 
 82 percent of those 60 and over. The same relationship between 
 age and increasing concern exists for crime. Also, females 
 expressed more concern about drugs and crime than did males. 
 
 The study we've been discussing was an in-home interview 
 situation. Now let me turn to a different kind of study completed 
 early in 1974 in San Diego with postcard interviews. We seldom 
 use this type of interview mechanism because in most cases it 
 
 Table 1. MOST IMPORTANT CITY PROBLEM 
 
 ( Percent ) 
 
 Problem 
 
 Aggregate 
 
 Downey 
 
 Bellflower 
 
 Lynwood 
 
 Paramount 
 
 Compton 
 
 Crime/police force 
 
 23 
 
 12 
 
 11 
 
 10 
 
 8 
 
 7 
 
 7 
 
 5 
 
 4 
 3 
 4 
 6 
 
 17 
 13 
 
 8 
 15 
 
 6 
 19 
 
 
 
 6 
 
 6 
 
 
 4 
 6 
 
 8 
 12 
 
 6 
 
 4 
 10 
 10 
 25 
 
 8 
 
 4 
 
 
 
 2 
 
 11 
 
 21 
 4 
 4 
 
 33 
 4 
 
 
 4 
 
 4 
 
 25 
 
 
 
 
 
 12 
 
 23 
 15 
 4 
 23 
 
 
 8 
 
 
 
 
 
 15 
 
 
 
 56 
 
 Pollution/noise 
 
 7 
 
 Teenagers/youth 
 
 21 
 
 Education 
 
 2 
 
 
 
 Shopping centers 
 
 
 
 9 
 
 Public/recreation facilities. 
 Pocketbook (taxes, inflation, 
 
 
 
 2 
 
 
 
 No problems 
 
 2 
 
 Other 
 
 7 
 
 
 
 Table 2. PUBLIC CONCERN OVER DRUG AND CRIME PROBLEMS 
 
 (Percent) 
 
 Characteristic 
 
 Drugs 
 
 Very 
 concerned 
 
 Moderately 
 concerned 
 
 Not 
 
 concerned 
 
 Crime 
 
 Very 
 concerned 
 
 Moderately 
 concerned 
 
 Not 
 
 concerned 
 
 Aggregate 
 
 Age 
 
 18 to 24 years 
 
 25 to 34 years 
 
 35 to 44 years 
 
 45 to 59 years 
 
 60 years and over. . . 
 
 Sex 
 
 Male 
 
 Female 
 
 City 
 
 Downey 
 
 Bellflower 
 
 Lynwood 
 
 Paramount 
 
 Compton 
 
 82 
 
 77 
 
 55 
 
 7fi 
 78 
 82 
 
 82 
 
 72 
 82 
 
 86 
 
 77 
 52 
 82 
 82 
 
 20 
 
 45 
 
 20 
 19 
 1(S 
 14 
 
 24 
 16 
 
 14 
 
 22 
 41 
 15 
 13 
 
 85 
 
 60 
 
 87 
 88 
 86 
 90 
 
 83 
 87 
 
 80 
 82 
 82 
 89 
 93 
 
 14 
 
 40 
 9 
 13 
 12 
 10 
 
 15 
 12 
 
 20 
 14 
 19 
 12 
 4 
 
THE ROLE OF PRIVATE FIRMS IN DEVELOPING INDICATORS 
 
 raises serious questions about how random the responses 
 may be. 
 
 Specifically, a mail questionnaire's validity rests on the kind 
 of response rates generated. To get valid results using this 
 technique you must do one of two things: either invest sufficient 
 resources to boost your response rate above 70 percent, or sample 
 the nonresponses to see if they were of different age, income, 
 or occupation groups, or whether they had different attitude sets 
 than those who did answer. 
 
 Then, what's the advantage of this technique? We inter- 
 viewed 5,000 people in San Diego at less cost than we interviewed 
 450 respondents in the Compton/Paramount areas because of the 
 sampling technique. Our initial recovery rate was above 50 
 percent without any followup letters. A sample of the nonre- 
 
 sponses confirmed that that population as a whole did match the 
 responding population. 
 
 This San Diego study illustrates the kind of survey research 
 data that can be generated with relatively little expense. Let's 
 review how the survey was conducted and then consider some of 
 the results. 
 
 The city manager sent the following letter (exhibit 1 ) to all 
 residents with their water bills. Note that the letter and the 
 interview were provided to respondents in both English and in 
 Spanish (exhibit 2). An article also appeared in the San Diego 
 Union which described the survey and said it would be used to 
 help the city evaluate community needs; undoubtedly this helped 
 improve the high recovery rate obtained for such a quick and 
 inexpensive survey. 
 
 EXHIBIT 1 
 
 We need your help. 
 
 OFFICE OF THE CITY MANAGER, CITY OF SAN DIEGO 
 
 The City of San Diego is conducting a survey among its citizens, and you 
 are one of several hundred who has been selected at random to participate 
 in this opinion poll. 
 
 Answering the questions on the enclosed questionnaire will take only 
 a few moments of your time. After you answer the questions, simply 
 detach the stub and return the large portion of the completed questionnaire 
 with your utility payment. There are no right or wrong answers, but your 
 response is very important to the accurancy of this survey. Your ideas 
 will be added to those of others and will give us a better understanding 
 of what the total community thinks. 
 
 We do not need to know your name. We do need your ideas and assistance. 
 Thank you in advance for helping us in this important task. 
 
 ^Mffvr^ 
 
 Kimball H. Moore , City Manager 
 
 QUESTIONNAIRE 
 
 THANK YOU 
 FOR YOUR HELP 
 
 If you have any questions 
 
 please call the 
 
 Comprehensive Management 
 
 Planning Program Office 
 
 at: 236-6060 
 
 I 1| Recycling solid wastes would conserve land. Would you be willing to set out cans 
 and bottles in a separate container for collection'' □ yes □ no 
 
 I 2 I For each recreation program listed below, please check those you feel should be 
 paid for by the City and those which should be paid for by the user: 
 
 RECORD BELOW 
 
 Paid for by City Paid for by User 
 Fresh Water Fishing Opportunities □ □ 
 Golf Programs □ □ 
 Athletic Programs □ □ 
 Aquatic Activities □ □ 
 ["51 Would you name two recreational activities or facilities most needed in the City of 
 San Diego2 / 
 
 ["41 Sex: □ MALE □ FEMALE 
 
 [31 Age: □ 18-24 □ 25-34 □ 35-44 □ 45-54 □ 55-64 □ 65 & OVER. 
 
 [g] Race: □ Caucasian □ Mexican American □ black □ oriental 
 
 □ other (Write in) 
 
 1/1 Home Ownership □ own home □ rent 
 
 ["81 Highest Grade of Formal Education: □ 1-8 □ 9-12 □ 1-3 yrs. college 
 
 □ 1-2 YRS VOCATIONAL SCH. □ 4 YRSCOLLEGE □ POST COLLEGE GRAD. 
 
 83 
 
THE ROLE OF PRIVATE FIRMS IN DEVELOPING INDICATORS 
 
 EXHIBIT 2 
 
 Estimados Residentes: 
 
 Necesitamos su ayuda. 
 
 La Ciudad de San Diego esta llevando a cabo un estudio entre sus residentes 
 que pagan agua y usted es una de las personas que hemos escogido para 
 participar. 
 
 Necesitar'a solamente unos minutos de su tiempo para contestar este cuestionario, 
 Simplemente despege la porcion mas chica y envie la porcion mas grande con su 
 pago de agua. Por favor no se le olvide incluir el cuestionario, puesto que 
 es muy importante para le exactitud del estudio. Sus opiniones junto con 
 otras opiniones nos diran lo que desea la comunidad entera. 
 
 No necesitamos su nombre solamente su ayuda e opiniones. 
 por su ayuda en este estudio importante. 
 
 Muchisimas gracias 
 
 /tyW^trttL^ 
 
 Kimball H. Moore Aministrador de la Ciudad 
 
 CUESTIONARIO 
 
 MUCHISIMAS GRACIAS 
 POR SU AUSILIO 
 
 Si tiene alguna pregunta 
 
 por favor, llame a la oficina 
 
 Comprehensive Management 
 
 Planning 
 
 al telefono 236-6060 
 
 1 1 1 El procesar despeirdisios sblidos ayudaria a conservar terrenos. i Esta usted 
 dispuesto a separar botes y botellas de su basura? □ si □ no 
 
 [~21 Favor de marcar su respuesta abajo de cada Programa Recreativa que debe ser 
 pagado por La Ciudad de San Diego o el usuario: 
 
 Pagado por la Ciudad Pagado por el usuario 
 
 □ □ 
 
 □ □ 
 
 □ □ 
 
 □ □ 
 
 donde pescar en agua dulce 
 
 Programas de Golf 
 
 Pro gramas de Atletismo 
 
 Programas Acuaticas 
 
 1 31 Nombre dos Programas Recreativos que asen falta en la Ciudad de San Diego: 
 
 L4J Sexo: □ MASCULINO □ FEMENINO 
 
 IS Edad: □ 18-24 □ 25-34 □ 35-44 □ 45-54 □ 55-64 □ 65 O MAYOR 
 
 |~6~1 Raza: □ blanca □ mexican-american □ negra □ asiatica 
 
 □ otras (indfquese) 
 
 1/1 Propiedad: □ dueno de su casa, o comprando □ renta su casa 
 |~oH Grado mas alto de educacion: □ 1-8 D 9-12 □ 1-3 anos de universidad 
 
 □ 1-2 ANOS DE ESCUELA VOCACIONAL □ ANOS DE UNIVERSIDAD 
 
 □ GRADUADO DE UNIVERSIDAD 
 
 Let's look briefly at some of the results that came out of 
 this little survey to demonstrate its value to public administrators. 
 From table 3 it is clear that the citizens identified a wide variety 
 of recreational needs. For example, in census subregions 1 and 2, 
 individual sports were most important. Supervised recreation was 
 very much desired in region 5— southeast San Diego. 
 
 In summary, as a result of the survey, the city of San Diego 
 had some idea of how best to allocate its scarce resources among 
 the various possible recreational activities in each region. 
 
 We also did a study in San Francisco designed to assess the 
 public's priorities for improving city services based on a willing- 
 ness to pay taxes. If you go out and ask people whether they 
 want a better environment, a better education for their children, 
 more effective police enforcement, or faster transporation, of 
 
 84 
 
 course, they'll say "yes" 99 percent of the time. But, if you ask 
 whether they're willing to spend more to have that service, even 
 to the point of increasing taxes, a much more realistic priority list 
 emerges. The San Francisco results are displayed in table 4. 
 
 Student: I think an interesting thing would be to isolate the 
 people that don't have children in school and examine their 
 spending priorities for the city to see if they would pay more 
 money for secondary and elementary education. 
 
 Mr. Wirthlin: Undoubtedly great differences among these types 
 of subgroups would result. Aggregate results don't tell the whole 
 or the most important side of the story. To utilize survey data 
 well you must break down the aggregate by family size, educa- 
 tion, income, location, etc. Only then can you begin to use 
 
THE ROLE OF PRIVATE FIRMS IN DEVELOPING INDICATORS 
 
 Table 3. RECREATION PROGRAM NEEDS BY CENSUS SUBREGION 
 
 ( Percent ) 
 
 Program 
 
 Parks/picnicking 
 
 Individual sports 
 
 Trails 
 
 Swimming pools 
 
 Supervised recreation 
 
 Ocean-bay activities 
 
 Camping 
 
 Team sports 
 
 Lake recreation 
 
 Senior citizens activities 
 
 'gregate 
 
 Census subregion 
 
 l(i 
 
 I I 
 
 13 
 
 11 
 
 i i 
 
 ■>:• 
 
 17.8 
 
 17.2 
 
 13.4 
 
 10.7 
 
 10.4 
 
 6.4 
 
 2.7 
 
 1.9 
 
 1.5 
 
 1.4 
 
 13.4 
 
 18.1 
 
 12.5 
 
 10.6 
 
 10.9 
 
 9.0 
 
 2.8 
 
 1.9 
 
 .6 
 
 2.8 
 
 16.5 
 
 21.4 
 
 9.9 
 
 11.0 
 
 11.0 
 
 7.1 
 
 3.8 
 
 1.1 
 
 2.2 
 
 .5 
 
 22.5 
 
 11.7 
 
 5.8 
 
 14.2 
 
 15.0 
 
 7.5 
 
 2.5 
 
 5.0 
 
 1.7 
 
 19.9 
 12.3 
 15.1 
 9.7 
 9.7 
 5.6 
 3.2 
 1.7 
 1.5 
 1.5 
 
 17.0 
 
 17.9 
 
 14.7 
 
 11.6 
 
 9.2 
 
 8.4 
 
 2.7 
 
 1.3 
 
 1.3 
 
 1.5 
 
 15.9 
 
 17.6 
 
 16.2 
 
 11.7 
 
 10.7 
 
 5.2 
 
 2.4 
 
 1.4 
 
 .7 
 
 1.0 
 
 17.0 
 
 23.2 
 
 11.6 
 
 10.3 
 
 10.3 
 
 3.1 
 
 3.1 
 
 1.3 
 
 1.8 
 
 .9 
 
 15.6 
 22. 2 
 16.0 
 
 6.6 
 12.7 
 
 4.2 
 .9 
 
 2.4 
 
 2.4 
 
 
 20.1 
 
 14.7 
 
 13.3 
 
 12.1 
 
 8.6 
 
 6.5 
 
 2. 1 
 
 2.7 
 
 2.1 
 
 1.2 
 
 37.8 
 
 15.6 
 
 2.2 
 
 8.9 
 
 11.1 
 
 4.4 
 
 2.2 
 
 4.4 
 
 
 
 2 2 
 
 Table 4. PRIORITIES OF SERVICES AND ACTIVITIES 
 BASED ON WILLINGNESS TO PAY 
 
 ( Percent ) 
 
 Program 
 
 Spend more 
 
 despite tax 
 
 increase 
 
 Spend more 
 if no tax 
 increase 
 
 Spend 
 less 
 
 Don' t 
 know 
 
 Cleaning up — protecting the 
 
 68.8 
 52.1 
 50.6 
 46.4 
 39.9 
 25.0 
 17.5 
 
 24.3 
 34.6 
 36.4 
 33.1 
 43.3 
 42.4 
 44.4 
 
 5.2 
 10.9 
 10.4 
 18.6 
 14.4 
 30.4 
 32.7 
 
 
 1 8 
 
 Elementary education 
 
 2 4 
 
 Secondary education 
 
 2 6 
 
 
 1 9 
 
 Public transportation 
 
 2 4 
 
 
 2 3 
 
 
 5.4 
 
 the data effectively as a policy tool. Once you understand where 
 the sources of support and strength for a given program lie, you 
 then can advance policy options that have a higher chance of 
 succeeding. 
 
 Student: What happens when you analyze that information and 
 find there are groups that don't want to spend more money for 
 education, or want to determine how and where the money is 
 spent? Making it public just reinforces the antagonism against a 
 lot of things that can't get through the legislature anyway, right? 
 
 Mr. Wirthlin: That's right. Information is power. How informa- 
 tion in a policy frame of reference is used depends on what your 
 objectives might be. For instance, one use of this study may have 
 been to help pass a school bond election. If that's the objective of 
 this study, it has one use. On the other hand it could have been 
 
 bought by taxpayer's groups who wanted to block further 
 bonding for schools, and for them it would have another use. Or, it 
 could have been funded by a board of supervisors who wanted to 
 know how the community ranked its own spending alternatives. 
 
 Student: Let's say I came to your firm and said, I want you to 
 find out some information that I could use against the powers 
 that be. Would you go out, understanding where I'm coming 
 from, and get the facts and figures to support me? 
 
 Mr. Wirthlin: Not necessarily. We don't accept as clients all those 
 who seek our services. 
 
 Student: What is your role once the survey results are compiled? 
 Are you in a position to act as a consultant? Do you shy away 
 from that or do you refer the client to someone else? 
 
 85 
 
THE ROLE OF PRIVATE FIRMS IN DEVELOPING INDICATORS 
 
 Mr. Wirthlin: I think the name of our firm. Decision Making 
 Information, defines our role. We provide individuals or groups 
 with information. They make decisions from it. Now, we can 
 partially bridge the gap between the data and its implementation 
 in policy by listing the options. But the client would have to say, 
 "Given my own set of interests, given my knowledge of the 
 community, given my assessment of the situation, this is what I 
 want to do." 
 
 Student: Could you give some more examples to illustrate the 
 scope and size of some of the survey projects? 
 
 Mr. Wirthlin: One of the larger studies that we completed 
 recently involved estimating gains from the Federal Government's 
 job training programs. We sampled, as I recall, over 5,000 trainees 
 in their homes and interviewed them again 12 months later. 
 
 Let me show you two tables generated from that study to 
 give you an idea of some of the things that were determined 
 about program participants. Naturally, we produced tables 
 showing demographic variables, such as sex, age, education, race, 
 ethnic background, marital status, etc. Other variables used in the 
 study included position in household, size of the household, 
 language spoken in the home, and welfare status (currently on 
 welfare) (table 5). Also, we measured some psychological 
 variables (table 6). 
 
 This study took about a year and a half to complete and 
 indicated that both on-the-job training and institutional training 
 substantially increased the incomes of those individuals who 
 completed the training, and also reduced unemployment among 
 this target group. 
 
 We've also done work with vocational training of the 
 handicapped. We have conducted studies for police agencies, 
 identifying their image and gauging their acceptance in the 
 community; we have worked for various community action 
 programs; for the Peace Corps; and we did some of the early work 
 on the new program of interviewing victims of crime. 
 
 Student: Do you think you might be doing something for the 
 Probation Department? 
 
 Mr. Wirthlin: We haven't worked directly in probation, but we 
 have done some work for law enforcement. In any one year we 
 have between 150 and 200 different projects. 
 
 Student: How does a person who needs research go about finding 
 the right group? 
 
 Mr. Wirthlin: There are a number of suppliers of this type of 
 research. If I were looking for a good survey research firm, I 
 would start with an academic recommendation or with a list of 
 survey research firms belonging to the American Marketing 
 Association or the American Association for Public Opinion 
 Research. I would select three or four from those listed and ask 
 each firm to supply me with a brief description of its experience, 
 sampling, and data analysis techniques. Then, to gain some 
 measure of their professional competence, I would also look at 
 the educational background of their staff. On the basis of this 
 information I would narrow my choice to two, and from these 
 request the names of some of their clients and determine if they 
 
 Table 5. HOUSEHOLD ENVIRONMENT OF JOB TRAINING PARTICIPANTS 
 
 
 Sample size 
 
 
 Proportional distribution 
 of sample 
 
 Environmental characteristics 
 
 (Number) 
 
 
 (Percent) 
 
 
 Institutional 
 
 On the 
 
 job 
 
 Institutional 
 
 On the job 
 
 Total 
 
 3,467 
 
 
 1,702 
 
 100 
 
 100 
 
 Position in household 
 
 
 
 
 
 
 
 1,940 
 1,527 
 
 
 910 
 792 
 
 56 
 44 
 
 54 
 
 Other 
 
 46 
 
 Size of household 
 
 
 
 289 
 590 
 736 
 642 
 470 
 740 
 
 
 102 
 282 
 387 
 355 
 238 
 338 
 
 8 
 17 
 21 
 19 
 14 
 21 
 
 6 
 
 Two 
 
 16 
 
 Three 
 
 23 
 
 
 21 
 
 Five 
 
 14 
 
 Six or more 
 
 20 
 
 Language spoken in household 1 
 
 
 
 3,394 
 
 
 1,684 
 
 98 
 
 99 
 
 
 407 
 
 
 155 
 
 12 
 
 9 
 
 Other 
 
 160 
 
 
 2 98 
 
 5 
 
 6 
 
 Welfare status 
 
 
 
 822 
 
 
 251 
 
 24 
 
 15 
 
 
 504 
 
 
 171 
 
 14 
 
 10 
 
 Never received welfare 
 
 2,141 
 
 
 1,280 
 
 62 
 
 75 
 
 
 
 'Use caution when interpreting results from these subgroups because of small number of sam- 
 ple cases. 
 
 2 Adds to more than total group because of bilingual households. 
 
 86 
 
THE ROLE OF PRIVATE FIRMS IN DEVELOPING INDICATORS 
 
 Table 6. PERSONAL OPINIONS OF JOB TRAINING PARTICIPANTS 
 
 
 Sample size 
 
 
 Proportional distribution 
 of sample 
 
 Opinion 
 
 (Number) 
 
 
 (Percent) 
 
 
 Institutional 
 
 On the 
 
 job 
 
 Institutional 
 
 On the job 
 
 Total 
 
 3,467 
 
 
 1,702 
 
 100 
 
 100 
 
 Value placed on work 
 
 
 
 
 
 
 High 
 
 2,692 
 
 
 1,339 
 
 78 
 
 79 
 
 
 555 
 220 
 
 
 275 
 J 88 
 
 16 
 6 
 
 16 
 
 
 5 
 
 Value placed on self-direction 
 
 
 
 
 
 
 High 
 
 1,198 
 
 
 585 
 
 34 
 
 34 
 
 Moderate 
 
 1,302 
 967 
 
 
 684 
 433 
 
 38 
 28 
 
 40 
 
 
 26 
 
 Levels of aspiration 
 
 
 
 
 
 
 High 
 
 602 
 
 
 385 
 
 17 
 
 23 
 
 Moderate 
 
 901 
 1,964 
 
 
 468 
 849 
 
 26 
 57 
 
 27 
 
 Low 
 
 50 
 
 
 
 1 Use caution when interpreting results from these subgroups because of small number of sam- 
 ple cases. 
 
 have been satisfied. Next, I would ask each to submit a proposal 
 related to my specific research task. 
 
 Student: How much time would you spend with me on a 
 proposal? 
 
 Mr. Wirthlin: It depends upon the potential size of the contract. 
 We develop proposals that cost $20.00; we have put out others 
 that have cost $5,000. Every client who has asked for a proposal 
 has been given one; it is the best check you can use when you go 
 into the market place to purchase survey research. 
 
 Student: How sophisticated are those working in the public 
 sector as far as understanding your methodology, approaches, 
 capabilities, and limitations? Do you find that you have to 
 educate people who seek your services? 
 
 Mr. Wirthlin: Over the last 10 years there's been some increase 
 in the knowledge people have about the use of attitude 
 measurement and survey research. Five years ago, I spent at least 
 half of my time explaining to people the use and usefulness of 
 doing survey research in the public sector. Today, I probably 
 spend less than 10 percent of my time in this activity. 
 
 Student: Is the employment training survey an example of one in 
 which you began with the null hypothesis? 
 
 Mr. Wirthlin: Yes. The null hypothesis was that training didn't 
 make any significant difference in income or employment. As it 
 turned out, it did. 
 
 Student: Now, that is different from the survey in the Para- 
 mount/Compton complex. That sounded more like a descriptive 
 study, where you just asked people what they thought. 
 
 Mr. Wirthlin: Yes, it was more descriptive, but we were gathering 
 data against the null hypothesis that pollution and not crime was 
 the burning issue of concern in the area. It turned out that the 
 data showed crime, not pollution, to be the major concern in the 
 specific communities. 
 
 In most cases, you begin by stating your objectives, what 
 you want to do. The survey researcher will work with you to set 
 up your null hypotheses, design the questionnaire, and create 
 analysis that will be responsive to the specific needs initially 
 identified. But you must know what you're really after; the 
 survey researcher can't identify your objectives for you. 
 
 Student: Has there been a tremendous proliferation of firms like 
 yourr in the last 10 years? 
 
 Mr. Wirthlin: Yes, but our industry remains one that is domi- 
 nated by 9 or 10 relatively large firms and a large number of 
 small, one- or two-man operations. 
 
 Student: In your proposal, do you state the research question 
 and the null hypothesis? 
 
 Mr. Wirthlin: Not always. Sometimes the null hypothesis is 
 implied. We begin our relationship with a client by saying, "If 
 you could answer the five most important questions in regard to 
 this project, what would they be?" He states the survey's 
 objectives in terms of his own questions. We aren't always able to 
 answer everyone's questions because some may be beyond the 
 realm of measurement. 
 
 Student. Is there any attempt within the industry to integrate 
 the kind of information that's generated or make it available 
 generally? For example, you might Go a particular study about 
 crime for a particular government agency in a certain area and 
 find that another overlapping county has contracted with a 
 
 87 
 
THE ROLE OF PRIVATE FIRMS IN DEVELOPING INDICATORS 
 
 separate organization to generate parallel or similar data, or 
 perhaps even conflicting data. Is there any attempt to integrate 
 that kind of information in either the public or the private 
 sector? 
 
 Mr. Wirthlin: The Federal Government provides a search service 
 called N.T.I.S. (National Technical Information Service) which 
 tracks all research done by the Government. For a relatively small 
 fee this agency will tell you what studies are available in your 
 areas of interest, and these may be purchased at nominal cost. 
 Also, DM I and other research firms donate a lot of information to 
 various university data banks. So, some is available through those 
 data banks. 
 
 Mr. Barabba: The Government also deals with this problem 
 through the Office of Management and Budget. As a government 
 agency, you cannot conduct a survey of more than nine people in 
 the United States unless you have an OMB clearance for it. The 
 purpose of this is to avoid asking duplicate questions of similar 
 groups. Also, if the questions had been asked before, and there 
 is some value to asking it in a similar way, it is the function of 
 OMB to insure proper coordination. 
 
 Student: With what kind of time factor can you operate without 
 sacrificing reliability? 
 
 Mr. Wirthlin: Last week we got a call on a Friday afternoon for a 
 particular study from a client who needed results by Tuesday. We 
 drew the sample, did the interviewing, computer and written 
 analyses within 3 days. However, on the average, a study 
 involving between 500 and 800 interviews would take about a 
 month from the time the client and DM I agree on the interview 
 schedule until we present results. 
 
 Student: What kind of interviewers do you use? Do you hire 
 professional interviewers? 
 
 Mr. Wirthlin: We hire professional interviewers, but we train 
 them specifically for each study. We put out a training manual on 
 our larger studies that explains the reasons for, and the subtleties 
 of, every question. The interviewers then go through a practice 
 interview that's tape recorded and played back to them with a 
 supervisor who corrects, advises, and encourages the interviewer 
 to insure that interviews are conducted in a uniform and proper 
 fashion. 
 
 Student: How do you estimate the cost for your client? 
 
 Mr. Wirthlin: The cost is determined principally by the size of 
 the sample, the length and complexity of the interview schedule 
 and the type of interview (in home, telephone, mail). We've done 
 enough of these so that we can estimate within 5 percent what 
 our cost will be and, in most cases, quote a fixed price. 
 
 Student: You say that your Los Angeles office does a lot of 
 national work. What do you do about interviewers? Do you take 
 them with you or do you recruit them at that particular site? 
 
 Mr. Wirthlin: We have supervisors in just about every area of the 
 United States. For a large study, the major supervisors from 
 Decision Making Information meet the field supervisors and are 
 present during the training session. It is the responsibility of the 
 field supervisors to see that the interview is conducted in the 
 proper fashion. 
 
 Let me close by reminding you of Bacon's statement, 
 "Knowledge is power. But mere knowledge is not power, it's only 
 possibility. Action is power and its highest manifestation is when 
 it's directed by knowledge." I hope you have gained some insight 
 into the particular kind of knowledge we call survey research so 
 that you can manifest whatever good intentions you have in 
 helping to build a better society. 
 
 88 
 
THE GRAPHIC PRESENTATION OF STATISTICS 
 
 Vincent P. Barabba, Director, U. S. Bureau of the Census 
 
 In a period of history such as ours, there can be little doubt that accurate, relevant, 
 timely, complete, and easily comprehensible data are essential. 
 
 Vincent Barabba 
 
 Note: During the first lecture each student was given a census 
 tract book of Flint, Mich., which was identified only as Center 
 City. Each was also provided with 10 blank census tract maps of 
 Center City and assigned the following problem discussed during 
 the presentation: 
 
 Remembering that each of you may represent a different 
 profession or sphere of interest, assume that your imme- 
 diate supervisor is going to Center City tomorrow, and that 
 he has asked to be provided with an overview of its most 
 relevant social and economic characteristics. In assigning 
 the project, he indicated that there would not be time for 
 him to read a detailed report. Therefore, you were in- 
 structed to prepare not more than 10 descriptive maps of 
 Center City to be used as background material during the 
 trip there. 
 
 The students prepared their maps based on the information 
 contained in the census tract book. Their maps displayed colored 
 tracts showing such varied information as median income, income 
 over $25,000, the black population, median education, age of 
 housing structures, and the number of families. 
 
 It was apparent that the number of possible factors to be 
 displayed would exceed the 10-map limit, and even 10 maps 
 would be cumbersome in reaching conclusions quickly. If only a 
 few maps were used, the information each supervisor desired 
 had to be clearly defined. However, Mr. Barabba pointed out that 
 one might employ a color-coded map, using certain highly mean- 
 ingful measurements, that provided persons, having a variety of 
 specific goals, a composite tool with which to make relevant judg- 
 ments. 
 
 Mr. Barabba: Our basic purpose in presenting data graphically is 
 to make it easier to use. Also, we hope to gain additional insight 
 into the practical aspects of data by displaying the tables in a 
 spatial setting. Then, we can see the interrelationships of 
 small-area characteristics. 
 
 There is, however, a basic trade-off involved in these 
 endeavors. First, the cost of graphic presentations is generally 
 more expensive than those of a tabular form. Second, a graphic 
 display's cost generally necessitates presenting only a portion of 
 the available data. So to some extent electing a graphics approach 
 represents a balancing between two important characteristics of 
 statistical presentation: (a) Accuracy and completeness, and (b) 
 ease of use. The following chart illustrates the dilemma. 
 
 Technique J 1 represents plotting a single variable on one 
 map. In that case the information is very easy to use, but it is 
 certainly not a complete community description, and it probably 
 will not be as accurate as a map containing several variables. 
 Technique T 5 represents the complete census tract book which is 
 very complete but may be extremely difficult to use, and the 
 
 accuracy of the conclusions drawn from the book may be 
 questionable for that reason. The ideal situation is technique T 6 
 representing ease of use without any loss of accuracy or 
 completeness. We want to avoid technique T 7 at all costs. 
 However, ideals are always difficult to obtain, and techniques T 2 , 
 T 3 , T 4 represent situations that fall between the extremes. There 
 are at least three factors that influence where a statistical 
 presentation falls within the scale: The importance of the 
 judgment to be made; how important accurate information is to 
 the quality of the judgment; and how much time is available to 
 make the decision. 
 
 ACCURATE 
 COMPLETE 
 
 ACCURACY 
 
 COMPLETENESS 
 
 CHARACTERISTIC 
 
 INACCURATE 
 INCOMPLETE 
 
 DIFFICULT ABILITY TO USE EASY 
 
 TO USE CHARACTERISTIC TO USE 
 
 Take the problem that was assigned to you. What were the 
 main points of the assignment? Your supervisor wants general 
 background characteristics of Center City; specific detail or com- 
 prehensiveness is not required. However, he does want it to be 
 accurate. Second, he does not have a lot of time to analyze the 
 results of the report. Third, you have no idea of the importance 
 of his initial assessment of Center City. 
 
 Let's review your assignment in the context of the trade-off 
 between accuracy and ease of use. We could go to the census tract 
 book that contains all the data on the census tracts of Center 
 City. When viewed as a whole, these data represent the key to 
 providing insight into certain complex phenomena about the city. 
 These data are quite complete, compared to the normally avail- 
 able data concerning cities. Certainly, there are benefits to using 
 
 89 
 
THE GRAPHIC PRESENTATION OF STATISTICS 
 
 tract data, although it is difficult and time consuming— a condi- 
 tion contrary to your supervisor's needs. The alternative to the 
 tract book approach is to develop a map format that is easy to 
 use; although, by its nature a graphic display cannot include 
 minute detail, and that's the trade-off. Given your supervisor's 
 need for an overview of Center City, less than complete informa- 
 tion might be acceptable. 
 
 This approach of graphically displaying data is exemplified 
 by a concept known as Social Area Analysis (SAA). 1 The basic 
 building blocks of SAA are drawn from Census Bureau tract 
 books. Depending on the purpose, the use of census data in its 
 published forms may present problems. Census information is 
 presented in a variety of ways, including census maps, congres- 
 sional district data books, the published results of the decennial 
 and economic censuses, electronic tapes, punchcards, and others. 
 Another problem is that the varied forms of census data are more 
 than matched by their volume. In fact, the decisionmaker's 
 dilemma can be summed up as possessing too much, rather than 
 too little, information. One need only thumb through the 
 published results of a census of housing and population to realize 
 how difficult it is to translate that mass of data into a useable 
 area information profile for decisionmaking. 
 
 You will recall from our first lecture the discussion of chart 
 5 (page 4) where I suggested that there is a relationship between 
 the degree of certainty a decisionmaker requires before he is 
 willing to commit to a decision and how much that certainty of 
 choice will cost in time and money. The SAA approach was 
 represented by the upper curve of chart 5, where selected 
 variables are calculated to contribute a greater degree of 
 explanation in a more concise manner than would otherwise be 
 possible. There are many other methods available to synthesize 
 data items into more useable indices, and I have selected SAA 
 because of its simplicity and, equally important, because it has 
 been empirically proven. 
 
 To accomplish the original goal of easily and accurately 
 describing a particular phenomenon, we need to determine the 
 fewest specific variables that will get us further along on the 
 easy-to-use scale without sacrificing accuracy. SAA organizes the 
 myriad of potential variables into classes which Shevky and Bell 
 believe accurately describe economic, family or ethnic character- 
 istics (fig. 1). But, to absorb even this ordered information can 
 prove quite difficult. Therefore, the information is further re- 
 duced to three basic indices as shown in figure 2. 
 
 Economic Status 
 
 The index of economic status is made up of two factors 
 selected because they accurately reflect many other factors. The 
 two elements that have been found to best describe economic 
 status are education and occupation. Income, for example, is not 
 used because income and education are generally related. There is 
 also the problem that a college professor may earn less money 
 than a truck driver. Therefore, economic status is not measured 
 
 'The original idea and the basic statistical technique from which SAA 
 was developed must be credited to Professors Eshref Shevky and Wendell 
 3ell and their outstanding work on Social Area Analysis. Those interested 
 in discussions of this technique should consult: Social Area Analysis, 
 Eshref Shevky and Wendell Bell (Stanford University Press: Stanford, 
 Calif., 1955), Stanford Sociological Series, Number 1 or Community 
 Structure and Analysis, Marvin B. Sussman, Editor (Thomas Y. Crowell 
 Company: New York, 1959). 
 
 FIGURE 1. 
 
 IMPORTANT VARIABLES 
 
 Total population 
 
 Race 
 
 Sex 
 
 Nativity 
 
 Married couples 
 
 Families or unrelated individuals 
 
 Number of dwelling units 
 
 Owner- or renter-occupied dwelling units 
 
 Type of structure 
 
 Condition and plumbing facilities 
 
 Year structure was built 
 
 Number of all occupied dwelling units 
 
 Number of persons in dwelling unit 
 
 Number of households 
 
 Population per household 
 
 Population in households 
 
 Institutional population 
 
 Years of school completed 
 
 Residence in 1955 
 
 Income 
 
 Age 
 
 Marital status 
 
 Employment status 
 
 Major occupation group 
 
 Women in the labor force 
 
 Persons per room 
 
 Type of heating fuel 
 
 Refrigeration equipment 
 
 Television 
 
 Monthly contract rent 
 
 Value of one-dwelling-unit structures 
 
 Spanish surnames (certain areas) 
 
 ORDERED 
 
 Economic Characteristics 
 
 Condition and plumbing facilities 
 
 Persons per room 
 
 Years of school completed 
 
 Income 
 
 Employment status 
 
 Major occupation group 
 
 Type of heating fuel 
 
 Refrigeration equipment 
 
 Monthly contract rent 
 
 Value of one-dwelling-unit structures 
 Family Characteristics 
 
 Sex 
 
 Married couples 
 
 Families or unrelated individuals 
 
 Owner- or renter-occupied dwelling units 
 
 Type of structure 
 
 Age 
 
 Marital status 
 
 Women in the labor force 
 
 Institutional population 
 Ethnic Characteristics 
 
 Race 
 
 Nativity 
 
 Spanish surnames 
 
 FIGURE 2. 
 Indices 
 
 Index of economic status 
 
 Education 
 
 Occupation 
 Index of familization 
 
 Fertility ratio 
 
 Women not in labor force 
 
 Owner-occupied dwelling units 
 Index of ethnic status 
 
 Race 
 
 Nativity 
 
 Spanish surnames (when available) 
 
 90 
 
THE GRAPHIC PRESENTATION OF STATISTICS 
 
 simply by the amount of money people earn. For our purposes, 
 economic status is a cross between the type of work and the 
 attainable goals or status that a good education implies. 
 
 We will soon discover that descriptions are not as accurate 
 when we reduce the number of variables included. While it is 
 much easier to use only two characteristics, the description is 
 really not as complete as if we used all those available. We have 
 created a definition of that status based on those subjective judg- 
 ments. This definition places us someplace along the scale be- 
 tween ease of use and completeness and accuracy within our 
 decisionmaker's matrix. If you accept as adequate my definition 
 of economic status, described by the person's job and education, 
 then we are further along the accuracy scale than if you argue 
 that income is just as important as occupation or education. Both 
 the provider of information and the user should be aware of this 
 trade-off as they attempt to synthesize reams of data into 
 easy-to-use formats. 
 
 In social area analysis, information concerning education is 
 secured by determining the ratio or proportion of the census tract 
 residents over 25 years of age who have less than an eighth-grade 
 education. Therefore, the greater the proportion of residents over 
 25 years of age with less than an eighth-grade education, the 
 lower the education score for the census tract. 
 
 Occupation information is secured by determining the ratio 
 or proportion of residents of a census tract who are classified as 
 laborers, operatives, or craftsmen, as opposed to secretarial, sales, 
 professional, etc. 2 That is to say, we develop a blue collar— white 
 collar ratio. Therefore, the greater the proportion of residents 
 who are classified as laborers, operatives or craftsmen, the lower 
 the occupation score for the census tract. 
 
 Familization Status 
 
 The term "familization" represents three ratios: Fertility, 
 women not in the labor force, and owner-occupied dwelling units. 
 These selected variables are intended to indicate changes in the 
 structure or function of elements within society at a level higher 
 than the immediate family. Fertility is intended to reflect differ- 
 ences in the relation of the population to the socioeconomic 
 system. And measures of housing type and women not in the 
 labor force should show differences in the function and structure 
 of the family. 
 
 Obviously, one characteristic of familization is children. 
 Consequently, we use a fertility ratio— the number of children 
 under 5 years of age in relation to the number of women 14 to 44 
 years, the child-bearing age span. Therefore, the greater the 
 proportion of children under 5 years of age in relation to women 
 14 to 44 years, the higher the familization score for that census 
 tract. 
 
 The second characteristic of familization is the mother's 
 role. No one would deny that the family situation of a working 
 mother is different than that of a nonworking mother. Therefore, 
 we use the measurement of women not in the labor force. The 
 greater the proportion of women not in the labor force, the 
 higher the familization score for that census tract. 
 
 The third characteristic of familization involves housing 
 information. Family situations differ according to the types of 
 
 housing units occupied, and historically it has been the case that 
 where persons owned their home the structure itself was that of a 
 single family dwelling unit. And, naturally enough, single family 
 units differ, to some extent, from multiple unit living in terms of 
 life style and social attitudes. For the 1960 tract book we are 
 synthesizing it was safe to assume that owner-occupied dwelling 
 units were single-family structures. However, with the rise of 
 condominium living we may need to alter our characteristics. 
 
 By combining these three factors, many different family 
 situations can be quickly and easily described. For example, a 
 high score of familization would describe a neighborhood which 
 is primarily made up of families having several children (fertility 
 ratio), living in single-family housing units (single-family dwelling 
 ratio), and with wives who generally remain at home (women not 
 in the labor force). 
 
 A possible index of ethnic status could be derived from 
 three factors: Race, nativity (country of origin of first- and 
 second-generation Americans), and Spanish surnames. (In the 
 censuses of 1960 and 1970, there was a special section developed 
 for Spanish surnames in the southwestern States. It is only 
 available for that section of the country.) In the interest of 
 brevity, we will not go into depth on the ethnic index, although 
 the same principles of graphic presentation would apply to it as 
 well. 
 
 Developing Indices 
 
 Let me next explain how SAA information may be 
 arranged for reading from computer printouts. As we read 
 through these printouts we'll focus on tracts 1 and 16 as our 
 examples. 
 
 Figure 3, "SAA Economic Index" consists of the occupa- 
 tion and education factors discussed earlier. First, look at the 
 occupation ratio. This column indicates that in tracts 1 and 16, 
 for every 1,000 employed people, 642.7791 and 166.5804 
 persons respectively, are classified as laborers, operatives, and 
 craftsmen. Under the education ratio column, for every 1,000 
 persons over 25 years of age, tract 1 has 434.2468 that have not 
 gone past the eighth grade and tract 16 contains 115.9570 that 
 have not. 
 
 The column titled "Std. Score" (standard score) represents 
 a simple concept. This technique permits us to place a great many 
 variables on a comparable scale by putting each geographic unit's 
 ratio for a specific variable on a scale of - 100. 3 This allows us 
 to view the relationship not only between the variables within 
 each geographic unit but also between the variables in different 
 units, using the standard score. However, this scale is not a mere 
 rank order. Rather, the computation places the given geographic 
 unit's ratio for a given factor in relationship to all other 
 geographic units in the area under study. 
 
 By adding together a given tract's standard scores and 
 dividing by the number of factors involved we derive the 
 economic index score. After each economic index score appears 
 the number 1, 2, 3, or 4, representing the division of the index 
 into four parts. The lowest portion, 0-24 is designated 1; 25-49 is 
 2; 50-74 is 3, and the highest, 75-100 is 4. 4 
 
 "Operatives" is a general classification of workers falling somewhere 
 below a craftsman in level of skill required and being namely those 
 who operate machinery excludingthose engaged in the operation of transport. 
 
 3 The formula for determining the standard score is detailed in Shevky 
 and Bell's works, op. cit. 
 
 4 We also could have divided the scale on the basis of the percent of 
 tracts that fell into various categories (e.g., top 25 percent or lowest 25 
 percent of all tracts). 
 
 91 
 
THE GRAPHIC PRESENTATION OF STATISTICS 
 
 When the occupation and education factors for tracts 1 and 
 16 are converted to standard scores, we find that tract 1, with an 
 occupation ratio of 642.7791, receives a standard score of 
 5.9324, while 16, with a ratio of 166.5804, receives a standard 
 score of 100. A score of 100 means that tract 16 has the fewest 
 number of residents classified as laborers, operatives, or craftsmen 
 of any census tract on the sheet, with the scoring arranged to 
 reflect the value judgment that the fewer the number of laborers, 
 operatives and craftsmen in a tract the higher the economic status 
 of that unit. 
 
 306.0747 children under 5 years. In the case of children the more 
 there are the more the tract is said to be familized, so the higher 
 the standard score. Thus, tract 1 receives a designation of 
 95.2145, while 16 is a 0.0000. 
 
 
 FIGURE 3. SAA ECONOMIC INDEX 
 
 
 Census 
 
 Occupation 
 
 Education 
 
 Economic 
 index 
 
 
 tract 
 
 Ratio 
 
 Std. score 
 
 Ratio 
 
 Std. score 
 
 
 1 
 
 642.7791 
 
 5.9324 
 
 434.2468 
 
 28.6723 
 
 17.302 
 
 -1 
 
 2 
 
 584.2244 
 
 17.4993 
 
 354.4380 
 
 46.5572 
 
 32.028 
 
 -2 
 
 3 
 
 658.7029 
 
 2.7869 
 
 495.2849 
 
 14.9938 
 
 8.890 
 
 -1 
 
 4 
 
 663.4189 
 
 1.8553 
 
 536.9531 
 
 5.6561 
 
 3.756 
 
 -1 
 
 5 
 
 576.6870 
 
 18.9882 
 
 562.1926 
 
 0.0000 
 
 9.494 
 
 -1 
 
 6 
 
 624.9143 
 
 9.4614 
 
 497.8335 
 
 14.4227 
 
 11.942 
 
 1 
 
 7 
 
 556.2031 
 
 23.0345 
 
 468.5586 
 
 20.9831 
 
 22.009 
 
 -1 
 
 8 
 
 366.8831 
 
 60.4325 
 
 418.6240 
 
 32.1733 
 
 46.303 
 
 -2 
 
 9 
 
 608.3105 
 
 12.7413 
 
 477.4692 
 
 18.9863 
 
 15.864 
 
 -1 
 
 10 
 
 463.2239 
 
 41.4015 
 
 291.4524 
 
 60.6720 
 
 51.037 
 
 -3 
 
 11 
 
 613.5962 
 
 11.6972 
 
 384.9197 
 
 39.7263 
 
 25.712 
 
 -2 
 
 12 
 
 574.9316 
 
 19.3349 
 
 493.2986 
 
 15.4390 
 
 17.387 
 
 -1 
 
 13 
 
 500.0000 
 
 34.1368 
 
 432.9158 
 
 28.9706 
 
 31.554 
 
 -2 
 
 14 
 
 507.3169 
 
 32.6915 
 
 438.1443 
 
 27.7989 
 
 30.245 
 
 -2 
 
 15 
 
 406.1392 
 
 52.6779 
 
 309.5544 
 
 56.6154 
 
 54.647 
 
 -3 
 
 16 
 
 166.5804 
 
 100.0000 
 
 115.9570 
 
 100.0000 
 
 100.000 
 
 -4 
 
 17 
 
 580.9480 
 
 18.1465 
 
 316.5193 
 
 55.0546 
 
 36.601 
 
 -2 
 
 18 
 
 611.9690 
 
 12.0186 
 
 400.1794 
 
 36.3067 
 
 24.163 
 
 -1 
 
 19 
 
 535.9609 
 
 27.0331 
 
 396.4377 
 
 37.1452 
 
 32.089 
 
 -2 
 
 20 
 
 496.1191 
 
 34.9034 
 
 428.5713 
 
 29.9441 
 
 32.424 
 
 -2 
 
 21 
 
 390.9312 
 
 55.6821 
 
 333.1924 
 
 51.3183 
 
 53.500 
 
 -3 
 
 22 
 
 289.7031 
 
 75.6785 
 
 223.9245 
 
 75.8048 
 
 75.742 
 
 ■1 
 
 23 
 
 476.5457 
 
 38.7700 
 
 370.0786 
 
 43.0522 
 
 40.911 
 
 -2 
 
 24 
 
 281.0515 
 
 77.3876 
 
 284.4827 
 
 62.2339 
 
 69.811 
 
 -3 
 
 25 
 
 458.3333 
 
 42.3676 
 
 350.7642 
 
 47.3805 
 
 44.874 
 
 -2 
 
 26 
 
 486.7161 
 
 36.7609 
 
 500.0000 
 
 13.9372 
 
 25.349 
 
 -2 
 
 27 
 
 380.6060 
 
 57.7217 
 
 301.8582 
 
 58.3401 
 
 58.031 
 
 -3 
 
 28 
 
 437.5962 
 
 46.4640 
 
 471.6636 
 
 20.2073 
 
 33.376 
 
 -2 
 
 29 
 
 395.9390 
 
 54.6929 
 
 364.4443 
 
 44.3148 
 
 49.504 
 
 -2 
 
 30 
 
 460.3499 
 
 41.9693 
 
 374.3147 
 
 42.1029 
 
 42.036 
 
 -2 
 
 31 
 
 557.8181 
 
 22.7155 
 
 455.8674 
 
 23.8272 
 
 23.271 
 
 -1 
 
 32 
 
 600.2764 
 
 14.3284 
 
 420.6763 
 
 31.7134 
 
 23.021 
 
 -1 
 
 33 
 
 5 97.0605 
 
 14.9636 
 
 447.7122 
 
 25.6547 
 
 20.309 
 
 -1 
 
 34 
 
 528.6755 
 
 28.4723 
 
 325.5303 
 
 53.0353 
 
 40.754 
 
 -■' 
 
 35 
 
 476.1511 
 
 38.8479 
 
 273.2798 
 
 64.7445 
 
 51.796 
 
 -3 
 
 36 
 
 345.6631 
 
 64.6243 
 
 269.3333 
 
 65.6289 
 
 65.127 
 
 -3 
 
 37 
 
 339.5249 
 
 65.8368 
 
 225.6201 
 
 75.4249 
 
 70.631 
 
 -:; 
 
 38 
 
 305.5554 
 
 72.5471 
 
 199.4060 
 
 81.2994 
 
 76.923 
 
 -4 
 
 39 
 
 404.9114 
 
 52.9205 
 
 247.3282 
 
 70.5601 
 
 61.740 
 
 -3 
 
 40 
 
 445.1997 
 
 44.9620 
 
 196.8162 
 
 81.8797 
 
 63.421 
 
 -3 
 
 41 
 
 402.1274 
 
 53.4704 
 
 298.3674 
 
 59.1224 
 
 56.296 
 
 -3 
 
 The education standard score designates tract 16 as a 100 
 score, while tract 1 receives 28.6723. This 100 score means that 
 of all the tracts, number 16 possessed the lowest proportion of 
 residents with less than an eighth-grade education. 
 
 When we combine the education and occupation standard 
 scores for tract 1 and divide by two, the number of factors, we 
 derive an economic index score of 17.302, which is rather low, 
 and the tract receives a number 1 designation, while tract 16's 
 combined and divided score receives an index score of 100 and a 
 number 4 designation. 
 
 Figure 4, the familization index, shows that under "fertility 
 ratio" for every 1,000 women between the ages of 14 and 44 in 
 census tract 1, there are 887.9446 children under 5 years of age, 
 while in tract 16 for every 1,000 designated women there are 
 
 
 
 Figure 4 
 
 . SAA FAMILIZATION INDEX 
 
 
 Census 
 tract 
 
 Fert 
 
 Llity 
 
 Women In 
 Labor Force 
 
 Single- family 
 dwelling 
 
 Family 
 status 
 
 
 Ratio 
 
 Std. score 
 
 Ratio 
 
 Std. score 
 
 Ratio 
 
 Std. score 
 
 index 
 
 1 
 
 887.9446 
 
 95.2145 
 
 314.4822 
 
 61.9785 
 
 780.2900 
 
 80.0612 
 
 7 9.0847 -A 
 
 2 
 
 775.5662 
 
 76.8254 
 
 340.1772 
 
 52.3170 
 
 800.4326 
 
 82.8156 
 
 70.6527 -B 
 
 3 
 
 796.7085 
 
 80.2851 
 
 336.7275 
 
 53.6141 
 
 592.2163 
 
 54.3434 
 
 62.7475 -B 
 
 4 
 
 667.5256 
 
 59.1462 
 
 302.4739 
 
 66.4936 
 
 521.7390 
 
 44.7060 
 
 5677819 -B 
 
 5 
 
 725.6279 
 
 68.6538 
 
 377.7239 
 
 38.1993 
 
 530.1606 
 
 45.8577 
 
 50.9036 -B 
 
 6 
 
 825.8601 
 
 85.0553 
 
 345.3872 
 
 50,3580 
 
 273.1543 
 
 10.7137 
 
 48.7090 -C 
 
 7 
 
 917.1892 
 
 100.0000 
 
 375.0000 
 
 39.2235 
 
 246.3884 
 
 7.0537 
 
 48.7591 -C 
 
 8 
 
 502.3696 
 
 32.1208 
 
 477.1321 
 
 0.8215 
 
 194.8051 
 
 0.0000 
 
 10.9808 -D 
 
 9 
 
 762.5730 
 
 74.6993 
 
 421.8750 
 
 21.5984 
 
 570.1243 
 
 51.3224 
 
 49.2067 -C 
 
 10 
 
 687.1333 
 
 62.3547 
 
 395.6184 
 
 31.4709 
 
 819.2295 
 
 85.3859 
 
 59.7372 -B 
 
 11 
 
 868.2205 
 
 91.9870 
 
 336.5442 
 
 53.6831 
 
 797.7161 
 
 82.4441 
 
 76.0380 -A 
 
 12 
 
 639.9490 
 
 54.6337 
 
 346.6577 
 
 49.8803 
 
 604.3955 
 
 56.0088 
 
 53.5076 -B 
 
 13 
 
 569.2598 
 
 43.0664 
 
 339.5889 
 
 52.5383 
 
 603.2195 
 
 55.8480 
 
 50.4842 -B 
 
 14 
 
 620.3472 
 
 51.4261 
 
 348.0391 
 
 49.3609 
 
 458.7412 
 
 36.0915 
 
 45.6262 -C 
 
 15 
 
 495.4441 
 
 30.9875 
 
 414.0127 
 
 24.5546 
 
 427.4753 
 
 31.8161 
 
 29.1194 -C 
 
 16 
 
 306.0747 
 
 0.0000 
 
 327.1570 
 
 57.2127 
 
 926.1016 
 
 100.0000 
 
 52.4042 -B 
 
 17 
 
 760.7449 
 
 74.4001 
 
 356.1965 
 
 46.2937 
 
 803.2520 
 
 83.2011 
 
 67.9650 -B 
 
 18 
 
 525.7729 
 
 35.9504 
 
 379.4253 
 
 37.5596 
 
 736.6816 
 
 74.0980 
 
 49.2027 -C 
 
 19 
 
 528.6819 
 
 36.4264 
 
 366.6956 
 
 42.3460 
 
 815.6565- 
 
 84.8973 
 
 54.5566 -B 
 
 20 
 
 524.1633 
 
 35.6870 
 
 345.7249 
 
 50.2311 
 
 702.5920 
 
 69.4365 
 
 51.7849 -B 
 
 21 
 
 446.4788 
 
 22.9751 
 
 392.9246 
 
 32.4838 
 
 693.9810 
 
 68.2590 
 
 41.2393 -C 
 
 22 
 
 504.8354 
 
 32.5243 
 
 370.1187 
 
 41.0589 
 
 710.4722 
 
 70.5141 
 
 48.0324 -C 
 
 23 
 
 539.3257 
 
 38.1681 
 
 367.7417 
 
 41.9527 
 
 599.5422 
 
 55.3451 
 
 45.1553 -C 
 
 24 
 
 548.1926 
 
 39.6191 
 
 365.4773 
 
 42.8041 
 
 589.4519 
 
 53.9653 
 
 45.4628 -C 
 
 25 
 
 578.8335 
 
 44.6330 
 
 405.7141 
 
 27.6749 
 
 509.5237 
 
 43.0357 
 
 38.4479 -C 
 
 26 
 
 670.4543 
 
 59.6254 
 
 312.1147 
 
 62.8686 
 
 393.3401 
 
 27.1483 
 
 49.8808 -C 
 
 27 
 
 440.6777 
 
 22.0258 
 
 479.3169 
 
 0.0000 
 
 523.1479 
 
 44.8987 
 
 22.3082 -D 
 
 28 
 
 496.1240 
 
 31.0988 
 
 355.1401 
 
 46.6909 
 
 224.8366 
 
 4.1066 
 
 27.2 988 -C 
 
 29 
 
 472.4263 
 
 27.2210 
 
 403.5242 
 
 28.4983 
 
 253.6585 
 
 8.0478 
 
 21.2557 -D 
 
 30 
 
 560.8108 
 
 41.6838 
 
 334.0908 
 
 54.6055 
 
 562.3440 
 
 50.2585 
 
 48.8493 -C 
 
 31 
 
 560.4551 
 
 41.6256 
 
 298.8762 
 
 67.8464 
 
 479.0286 
 
 38.8657 
 
 49.4459 -C 
 
 32 
 
 594.4517 
 
 47.1887 
 
 338.1294 
 
 53.0870 
 
 769.6077 
 
 78.6004 
 
 59.6254 -B 
 
 33 
 
 615.9272 
 
 50.7029 
 
 318.5408 
 
 60.4524 
 
 527.7410 
 
 45.5268 
 
 52.2273 -B 
 
 34 
 
 600.5 933 
 
 48.1937 
 
 328.4534 
 
 56.7252 
 
 844.8740 
 
 88.8926 
 
 64.6038 -B 
 
 35 
 
 622.5264 
 
 51.7827 
 
 314.6943 
 
 61.8987 
 
 888.7202 
 
 94.8883 
 
 69.5232 -B 
 
 36 
 
 443.4468 
 
 22.4789 
 
 367.1304 
 
 42.1825 
 
 888.8887 
 
 94.9113 
 
 53.1909 -B 
 
 37 
 
 563.6018 
 
 42.1405 
 
 363.8496 
 
 43.4161 
 
 910.7791 
 
 97.9047 
 
 61.1538 -B 
 
 38 
 
 509.6582 
 
 33.3135 
 
 288.9316 
 
 71.5856 
 
 903.9546 
 
 96.9715 
 
 67.2902 -B 
 
 39 
 
 518.8582 
 
 34.8189 
 
 325.6519 
 
 57.7786 
 
 919.5906 
 
 99.1096 
 
 63.9024 -B 
 
 40 
 
 830.6218 
 
 85.8345 
 
 302.1226 
 
 66.6257 
 
 919.1313 
 
 99.0468 
 
 83.8357 -A 
 
 41 
 
 516.8809 
 
 34.4953 
 
 379.0847 
 
 37.6877 
 
 716.9810 
 
 71.4041 
 
 47.8624 -C 
 
 In relation to working women (figure 4) there are 314.4822 
 women in the labor force for every 1,000 women over 14 years of 
 age in census tract 1, and 327.1570 in tract 16. Tract 1 receives a 
 standard score of 61.9785 and 16 a score of 52.2127. 
 
 The last factor is owner-occupied dwelling units with tract 
 1 having a ratio of 780.2900 and receiving a standard score of 
 80.0612 and 16 a ratio of 926.1016 and a score of 100. 
 
 As in the case of the "Economic Index," we add a given 
 tract's standard scores together and divide by the number of 
 factors, three in this case, to obtain the designation for the index 
 of family status. Again, the 0-100 index is divided into parts 
 identical to the Economic Index, but the designations are now 
 alphabetic with 0-24 the lowest portion, labeled D, and the 
 highest portion, 75-100 labeled A. 
 
 Thus, for the purposes of family status tract 1 is an A tract, 
 while tract 16 is B. 
 
 If, at this point, you feel this technique is quite compli- 
 cated, imagine the complexity of having to use 19 variables 
 instead of the 5 we employed in the economic and familization 
 indexes. 
 
 All that we have done is describe four levels of economic 
 status— 1 being low and 4 being high— on the basis of occupation 
 and education. Next, we have described four levels of familization— 
 D being low and A being high— on the basis of number of 
 children, women in the labor force, and single-family dwelling 
 units. 
 
 92 
 
THE GRAPHIC PRESENTATION OF STATISTICS 
 
 Next we can further reduce the data by combining these 
 levels of status into one easy-to-read table (fig. 5). In one glance 
 the table allows us to see the overall socioeconomic profile of any 
 census tract. For example, in tract 1 we have an economic index 
 designation of 17.302 or a 1, and an index of family status of 
 79.0847 or A. Therefore, the SAA score of census tract 1 is 1-A. 
 
 Figure 5. Social Area Key 
 
 Census 
 
 Economic 
 
 Economic 
 
 Family 
 
 Family 
 
 tract 
 
 index 
 
 rank 
 
 rank 
 
 index 
 
 1 
 
 17.302 
 
 1 
 
 A 
 
 79.0847 
 
 2 
 
 32.028 
 
 2 
 
 B 
 
 70.6527 
 
 3 
 
 8.890 
 
 1 
 
 B 
 
 62.7475 
 
 4 
 
 3.756 
 
 1 
 
 B 
 
 56.7819 
 
 5 
 
 9.494 
 
 1 
 
 B 
 
 50.9036 
 
 6 
 
 11.942 
 
 1 
 
 C 
 
 48.7090 
 
 7 
 
 22.009 
 
 1 
 
 C 
 
 48.7591 
 
 8 
 
 46.303 
 
 2 
 
 D 
 
 10.9808 
 
 9 
 
 15.864 
 
 1 
 
 C 
 
 49.2067 
 
 10 
 
 51.037 
 
 3 
 
 B 
 
 59.7372 
 
 11 
 
 25.712 
 
 2 
 
 A 
 
 76.0380 
 
 12 
 
 17.387 
 
 1 
 
 B 
 
 53.5076 
 
 13 
 
 31.554 
 
 2 
 
 B 
 
 50.4842 
 
 14 
 
 30.245 
 
 2 
 
 C 
 
 45.6262 
 
 15 
 
 54.647 
 
 3 
 
 C 
 
 29.1194 
 
 16 
 
 100.000 
 
 4 
 
 B 
 
 52.4042 
 
 17 
 
 36.601 
 
 2 
 
 B 
 
 67.9650 
 
 18 
 
 24.163 
 
 1 
 
 C 
 
 49.2027 
 
 19 
 
 32.089 
 
 2 
 
 B 
 
 54.5566 
 
 20 
 
 32.424 
 
 2 
 
 B 
 
 51.7849 
 
 21 
 
 53.500 
 
 3 
 
 C 
 
 41.23 93 
 
 22 
 
 75.742 
 
 4 
 
 C 
 
 48.0324 
 
 23 
 
 40.911 
 
 2 
 
 C 
 
 45.1553 
 
 24 
 
 69.811 
 
 3 
 
 C 
 
 45.4628 
 
 25 
 
 44.874 
 
 2 
 
 C 
 
 38.4479 
 
 26 
 
 25.349 
 
 2 
 
 C 
 
 49.8808 
 
 27 
 
 58.031 
 
 3 
 
 D 
 
 22.3082 
 
 28 
 
 33.376 
 
 2 
 
 C 
 
 27.2988 
 
 29 
 
 49.504 
 
 2 
 
 D 
 
 21.2557 
 
 30 
 
 42.036 
 
 2 
 
 C 
 
 48.8493 
 
 31 
 
 23.271 
 
 1 
 
 C 
 
 49.4459 
 
 32 
 
 23.021 
 
 1 
 
 B 
 
 59.6254 
 
 33 
 
 20.309 
 
 1 
 
 B 
 
 52.2273 
 
 34 
 
 40.754 
 
 2 
 
 B 
 
 64.6038 
 
 35 
 
 51.796 
 
 3 
 
 B 
 
 69.5232 
 
 36 
 
 65.127 
 
 3 
 
 B 
 
 53.1909 
 
 37 
 
 70.631 
 
 3 
 
 B 
 
 61.1538 
 
 38 
 
 76.923 
 
 4 
 
 B 
 
 67.2902 
 
 39 
 
 61.740 
 
 3 
 
 B 
 
 63.9024 
 
 40 
 
 63.421 
 
 3 
 
 A 
 
 83.8357 
 
 41 
 
 56.296 
 
 3 
 
 C 
 
 47.8624 
 
 If we were to describe the economic and familization dif- 
 ferences between tracts 1 and 16 based on the Social Area Key 
 we could say that residents in tract 1 tend to be laborers, opera- 
 tives, and craftsmen, with less than an eighth-grade education; 
 while residents of tract 16 tend to work in capacities other than 
 those prevalent in tract 1, such as secretarial, sales and profes- 
 sional, and tend to be educated beyond the eighth-grade level. 
 
 As regards familization differences, we know that for 
 "family rank", on the Social Area Key, tract 1 scores a 79.0847, 
 placing it in the highest category and receives an A designation. 
 While tract 16 scores only 52.4042, making it a B. In reviewing 
 the two tract's familization scores we discover that the standard 
 scores for women in the labor force and owner-occupied 
 dwellings are not significantly different. In terms of familization 
 the only significant difference is the number of children under 5 
 years of age in relation to women between the ages of 14 and 44 
 years. When the three standard scores are combined and divided, 
 naturally, census tract 1 receives a higher index of family status 
 because of its higher fertility standard score. 
 
 Therefore, if we were describing the familization differences 
 between the two census tracts, we would say that residents in 
 census tract 1 tend to have many young children in their families. 
 The women tend not to hold jobs, and families usually own their 
 housing units. Residents of tract 16 almost always reside in 
 owner-occupied units, the women tend not to hold jobs, and 
 there are usually no children under 5 years of age in that tract's 
 households. 
 
 Based on our description of the two tracts it is possible to 
 assume that the resident's needs and concerns may differ 
 considerably, and a recognition of this difference and the factors 
 which cause it can be of crucial importance for effectively dealing 
 with each tract's problems. 
 
 A method of illustrating the relationships between eco- 
 nomic and familization status and highlighting the differences 
 between census tracts is to construct a two-variable matrix. 
 
 In Figure 6, Economic Index Key, there are four boxes. 
 
 FIGURE 6. 
 
 ECONOMIC INDEX KEY 
 
 On the other hand, census tract 16 has an economic index of 100, 
 or 4, and an index of family status of 52.4042, or B. Thus tract 
 16 has an SAA score of 4-B. 
 
 1 
 
 2 
 
 3 
 
 4 
 
 25 
 
 50 
 ECONOMIC 
 
 75 
 
 100 
 
 93 
 
THE GRAPHIC PRESENTATION OF STATISTICS 
 
 From left to right, each box represents a range of the to 100 
 economic standard scores (box 1 = to 24, box 2 = 25 to 49, box 
 3 = 50 to 74, and box 4 = 75 to 100). 
 
 On the other hand, in Figure 7, Familization Index Key, 
 the boxes, A, B, C, and D, represent the range of the familization 
 standard scores. (Running from bottom to top: Box D = to 24, 
 box C = 25 to 49, box B = 50 to 74, and box A = 75 to 1 00.) 
 
 FIGURE 7. 
 100 
 
 FAMILIZATION INDEX KEY 
 
 75 
 
 2 
 O 
 
 H 
 < 
 H 50 
 
 25 
 
 B 
 
 D 
 
 FIGURE 8. 
 100 
 
 75 
 
 < 
 N 50 
 
 25 
 
 1A 
 
 2A 
 
 3A 
 
 4A 
 
 1B 
 
 2B 
 
 3B 
 
 4B 
 
 1C 
 
 2C 
 
 3C 
 
 4C 
 
 1D 
 
 2D 
 
 3D 
 
 4D 
 
 25 
 
 50 
 ECONOMIC 
 
 75 
 
 100 
 
 When the two indices are combined into the social area 
 typology, figure 8, it can be seen that there are 16 different types 
 of census tracts based on economic and familization status. 
 
 94 
 
 Now, how do we relate the 41 census tracts of Center City 
 to the social area typology? Actually, it's quite simple (fig. 9). 
 
 FIGURE 9. SOCIAL AREA INDEX OF CENTER CITY 
 
 100 
 95 
 90 
 85 
 80 
 75 
 70 
 
 : 65 
 
 60 
 
 N 45 
 
 f 40 
 
 < 
 "■ 35 
 
 30 
 
 25 
 
 20 
 
 15 
 
 10 
 5 
 
 
 
 1 1 
 
 I I 
 
 1 1 
 2A 
 
 1 1 1 1 
 3A 
 
 1 
 
 1 1 
 4A- 
 
 • i 
 
 1 
 
 
 • 
 
 
 - 
 
 1B 
 • • 
 
 • 
 • 
 
 2B 
 
 • 
 
 . - 
 
 • 
 • • 
 
 » 
 
 4B 
 
 *• # v* 
 
 
 • 
 
 
 © 
 
 1C 
 
 • 
 
 ; 2C 
 
 • 
 
 a 3C 
 
 • • 
 
 • 
 
 4C_ 
 
 - 
 
 • 
 
 
 • 
 
 
 - 
 
 1D 
 
 
 2D( 
 
 1 • 3D 
 
 
 4D 
 
 _ 
 
 
 • 
 
 
 
 _ 
 
 I I I I 
 
 l l 
 
 1 1 
 
 1 1 1 1 
 
 1 I 
 
 ' 1 
 
 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 
 ECONOMIC STATUS 
 
 We have placed the 100-point economic index on a hori- 
 zontal scale and the 100-point familization index on a vertical 
 scale. Let's take census tract 16 again. The economic score of 
 census tract 16 is 100; so we move to 100 on the economic index 
 (horizontal scale). Tract 16 scores 52.4 on the familization index 
 (vertical axis) so we move up to 52 and place the number 16 at 
 that point. Additionally, we know that 100 equals 4 in economic 
 status and 52.4 equals B in familization status; therefore, census 
 tract 16 is a 4-B census tract. 
 
 At this point we have taken our census tract book of Center 
 City and reduced it to one typology, based on five characteristics, 
 which relates the economic and familization status of all the 
 census tracts in Center City. Clearly, we have moved quite a way 
 on our scale towards "easy to use." In fact, we can now quantify 
 how far we have moved along the scale. Let me briefly review the 
 exact extent to which we have condensed the original tract data. 
 First, we took our census tract book with all of its 32 detailed 
 characteristics of each of the 41 census tracts creating 1,312 data 
 items. We made some subjective judgments and decided to use 
 only those variables that social scientists have used in previous 
 studies. This activity provided us with two key categories of data 
 (actually three categories when ethnicity is utilized): economic 
 and familization status. At that point we were working with 10 
 economic variables and 9 familization variables in each of our 41 
 census tracts— in all, 779 data items. 
 
 1 economic variables 
 9 familization variables 
 
 x 41 census tracts 
 x 41 census tracts 
 
 410 data items 
 369 data items 
 
 19 variables 
 
 x 41 census tracts 
 
 779 data items 
 
 We systematically determined which variables in our two 
 concepts most efficiently described each of our concepts. We 
 found that economic status could be described by education and 
 occupation characteristics, and that familization status could be 
 
THE GRAPHIC PRESENTATION OF STATISTICS 
 
 described by fertility, working women, and home ownership char- 
 acteristics. By this further reduction we were now dealing with 
 two economic variables and three familization variables in each of 
 our 41 census tracts— in all 205 data items. 
 
 VARIABLE 
 
 STD. SCORE 
 
 2 economic variables 
 
 3 familization variables 
 
 41 census tracts 
 41 census tracts 
 
 82 data items 
 123 data items 
 
 5 variables 
 
 41 census tracts 
 
 205 data items 
 
 We normalized these variables within each concept by 
 placing them on a standard score scale of from to 100. From 
 this procedure we derived one index score for each of our two 
 concepts in each of our 41 census tracts— in all 82 data items. 
 
 1 economic index 
 1 familization index 
 
 41 census tracts 
 41 census tracts 
 
 41 data items 
 41 data items 
 
 2 indices 
 
 41 census tracts 
 
 82 data items 
 
 Finally, by collapsing the 100-point indices into four cate- 
 gories and assigning alphabetic and numeric designations, we 
 combined the four categories for each concept and created 16 
 possible positions within a matrix, allowing us to place each of 
 our 41 census tracts in a "social space" describing economic and 
 familization relationships. 
 
 In developing this easy-to-use typology, have we, in fact, 
 traded off accuracy? The answer is yes and no. Let's look first at 
 the economic status index of census tract 16 (fig. 3). We see that 
 census tract 16 ranks highest in economic status with a score of 
 100. This implies, if we have not lost any accuracy, that census 
 tract 16 should rank among the highest tracts in all of our eco- 
 nomic variables. We know already, in the case of our two "most 
 efficient" variables, occupation and education, that it ranks high- 
 est. But what about the other variables? Does our economic 
 status index accurately really reflect them? The census tract data 
 book contains the information necessary to verify our theory. 
 Plotting all the relevent economic characteristics of tract 16 
 would have yielded the following results: 
 VARIABLE RANK 
 
 Income 1st 
 
 Persons per room 1st 
 
 Type of heating fuel 2nd 
 
 Contract monthly rent No Renters 
 
 Value of owner-occupied dwellings 1st 
 
 Condition and plumbing facilities 2nd 
 
 In fact, for census tract 16 we have developed an index which is 
 easy to use and one in which we have lost very little accuracy, 
 while eliminating unnecessary and cluttering detail. In other 
 words, the economic index is an excellent index of economic 
 status in Center City. 
 
 Let's look now at familization status. The index is 
 52.4— almost the median point (23 tracts below and 17 tracts 
 above). However, a review of the standard scores of the elements 
 which contributed to the index number 52.4 reveals that census 
 tract 16 is not at all in the middle, relative to its familization 
 variables. In this case, census tract 16 has the lowest fertility ratio 
 and the highest owner-occupied dwelling unit ratio and is in the 
 highest third of all tracts relative to women not working. Yet, 
 when we look only to The index, we are left with the impression 
 that census tract 16 is about at the mid-point of familization 
 status. Clearly, in this case, we have lost accuracy as we moved 
 toward the easy-to-use side of our scale. 
 
 Fertility 0.0 
 
 Women not in labor force 57.2 
 
 Single-family dwelling units 100.0 
 
 RANK 
 
 Lowest 
 
 .Third highest 
 
 Highest 
 
 1 57.2 t 3=52.4 
 
 From a decisionmaker's point of view, however, it is impor- 
 tant to note that there was sufficient detail in the presentation of 
 figure 4 to reflect that deficiency in the index. This example 
 highlights the responsibility of the provider of information to 
 present the decisionmaker with sufficient, detailed information 
 and methodological explanation to enable him to determine how 
 much confidence he can place in the data. 
 
 Although we have certainly reduced the volume of data 
 items to manageable proportions, we haven't yet made our super- 
 visor's map. Remember, he said he wanted an overview of Center 
 City in no more than 10 maps. Well, since he's so busy, let's see if 
 we can do it all in one map. 
 
 To accomplish that we will use the three primary colors- 
 red, blue, and yellow— to relate the two concepts of economic 
 and familization status. Let me first illustrate how these primary 
 colors relate to the color spectrum. 
 
 If we take two vertical bars of yellow and red, and blend 
 them with two horizontal bars of yellow and blue we will pro- 
 duce the four squares of purple, orange, green and yellow (fig. 10). 
 
 FIGURE 10. COLOR-CODED CONCEPT 
 
 
 Using the same colors, but, in this case separating the verti- 
 cal red bar into three shades of red and the horizontal blue bar 
 into three shades of blue, we blend the vertical and horizontal 
 bars in the same way, creating a four by four matrix of 16 squares 
 of different colors and shades. In the corners of our matrix we 
 find the same purple, orange, green and yellow colors that we 
 generated when we blended the two vertical and horizontal bars 
 of primary colors. The basic difference between the 4-color 
 matrix and the 16-color matrix is the tones that are created when 
 we shade the blue and red (fig. 11). 
 
 FIGURE 
 
 COLOR-CODED MATRIX 
 
 
 
 
 
 ■ ■ ' ■ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 Now, how do we relate this color-coding technique to the 
 map we must make for Center City? Lets go back to figures 6 and 
 7 where we developed the economic index key and the familiza- 
 tion index key by dividing their respective indices into quadrants. 
 In figure 12 we use the same index keys but we color-code them. 
 
 95 
 
THE GRAPHIC PRESENTATION OF STATISTICS 
 
 FIGURE 12. COLOR-CODED SOCIAL AREA INDEX 
 100 
 
 A 
 
 75 
 
 < 
 N 50 
 
 < C 
 
 25 
 
 1A 
 
 2A 
 
 3A 
 
 
 1B 
 
 2B 
 
 3B 
 
 4B 
 
 1C 
 
 2C 
 
 3C 
 
 4C 
 
 1D 
 
 2D 
 
 3D 
 
 4D 
 
 MAP 2. CENTER CITY FAMILIZATION PROFILE 
 
 J I 4J II Ks I L 
 
 25 
 
 2 50 3 
 
 ECONOMIC 
 
 75 
 
 100 
 
 MAPI. CENTER CITY ECONOMIC PROFILE 
 
 In map 1 we take the lowest economic status (1) and color 
 it yellow. The second lowest (2) is color-coded light red, the 
 second highest (3) is color-coded medium red and the highest 
 economic status (4) is color-coded dark red. In map 2 we take the 
 lowest familization (D) and color-code it yellow. The second 
 lowest (C) is color-coded light blue, the second highest (B) is 
 color-coded medium blue, and the highest familization status (A) 
 is color-coded dark blue. These two maps show us where the 
 different types of persons, as identified by their economic and 
 
 MAP 3. CENTER CITY SAA PROFILE 
 
 familization status, are located in Center City. As far as economic 
 status is concerned, the higher economic census tracts are located 
 north of the river and clustered together. As far as familization 
 status is concerned, the lowest familization is located in the 
 center of the city, and as you develop concentric circles further 
 from center city, familization status gets higher. 
 
 So, now we have two maps which should provide your 
 supervisor with an excellent overview of Center City. And, if you 
 wish, we can blend the two maps together creating map 3, which. 
 
 96 
 
THE GRAPHIC PRESENTATION OF STATISTICS 
 
 in one map, provides an excellent overview for your supervisor's 
 use. 
 
 The Census Bureau's geography staff has recently developed 
 the ability to produce full-color, cross-classification maps using 
 computerized mapping techniques and COM (Computer Output 
 on Microfilm) units. Because of the reduction in costs and time, 
 this new map-making ability opens up a variety of analytic uses 
 not before feasible. 
 
 Now let's see how this relates to relevant census informa- 
 tion. In map 4 we have color-coded, using red and yellow, the 
 average size of the farms in the 3,000 counties within the contigu- 
 ous 48 States. Yellow indicates those counties where the average 
 size of farms is less than 50 acres. The next classification is repre- 
 sented by light red and illustrates counties where farms average 
 50 to 199 acres. Medium red represents counties where farms 
 average 200 to 499 acres. The dark red areas are counties where 
 farms average 500 or more acres. As you can see, although there 
 is great variety in the size of American farms, there tends to be a 
 geographic clustering of the different-size farms. The large yellow 
 area in the mid-section of the eastern United States contains over 
 half of all the Nation's farms of 50 acres or less. Most of these 
 farms are family farms operated by part-time farmers. The larger 
 farms, particularly large wheat farms, some as large as 2,000 
 acres, are found in the midwest and western parts of our country. 
 
 Map 5 shows the average value of all agricultural products 
 sold per farm in each of the 3,000 counties. The yellow indicates 
 counties where farms average under $5,000 in the value of prod- 
 ucts sold. The light blue represents average sales of $5,000 to 
 $19,999. The medium blue represents counties with average sales 
 of $20,000 to $39,999, and the dark blue reflects average sales of 
 $40,000 or more. 
 
 At this point, one might hypothesize that the dark blue 
 areas, representing the highest average value of products sold per 
 farm, would tend to be located in the dark red areas, representing 
 the largest farms. Additionally, we would not expect to find 
 green areas which would be the combination of blue (high average 
 values) with yellow (small size) or orange areas, which would be 
 the combination of red (large size) with yellow (low average 
 values). 
 
 How does our hypothesis stand the test? Well, first, map 6 
 clearly illustrates that the size of farm and the average value of 
 agricultural produce sold are not necessarily related. In fact, there 
 are many areas of our country (represented by green) in which 
 the smaller farms (yellow) have high average values of product 
 sold (blue). In fact, the products of the small farms— vegetables in 
 the northeast around New York City, specialized orchards in the 
 Michigan area, and tobacco in North Carolina— rather than the 
 sizes of the farms, affect average value in dollars. 
 
 Using computer output and microfilm plotting devices, the 
 Census Bureau can produce these types of maps quickly and in- 
 expensively. The low cost and short preparation time of these 
 new techniques now make it possible for the research analyst to 
 prepare many graphic displays to test the validity of various data 
 relationships, or more clearly identify particular geographic areas 
 of interest or concern. Our eventual goal is to make this mapping 
 technique available at the census tract level, as well as the county 
 level. This next set of charts is an example of the maps which will 
 be produced in our forthcoming Urban Atlas. 5 
 
 In map 7 we have color-coded the census tracts of the Fort 
 Worth SMSA according to the median income of all the families 
 in each tract. The yellow area represents census tracts of $7,000 
 income or less; the light red, $7,000 to $9,999; the medium red, 
 $10,000 to $14,999; the dark red, $15,000 or more. 
 
 Map 8 shows the percentage of the adult population which 
 has graduated from high school. The yellow area represents tracts 
 where less than 35 percent of the adults have graduated from high 
 school; the light blue represents 35 to 64.9 percent; the medium 
 blue, 65 to 84.9 percent; and the dark blue, 85 percent or more 
 of the adults have completed 12 or more years of education. The 
 blending of these two maps (map 9) illustrates the relatively 
 strong correlation between median family income and completion 
 of 12 years of education. 6 However, there are green and orange 
 areas in which the percent of high school graduates does not seem 
 to affect the median family income. From an analytic point of 
 view, these might merit further investigation to determine why 
 these two characteristics are not related in those areas. 
 
 In conclusion, I want to stress that the concepts we have 
 discussed are preliminary, and there is still a tremendous potential 
 to be developed in the graphic presentation of statistics. For ex- 
 ample, in this presentation I have made no mention of the 
 psychological aspects of presenting graphics, another aspect still 
 in its fledgling stages. At this point, we have little knowledge of 
 how often decisionmakers misinterpret graphic displays, or how 
 often such presentations are rejected out of hand. However, in a 
 period of history such as ours, there can be little doubt that 
 accurate, relevant, timely, complete, and easily comprehensible 
 data are essential. I would respectfully submit that our discussion 
 affords an overview of the unique possibilities offered by graphic 
 displays. 
 
 5 Urban Atlas, Series GE-80, U.S. Bureau of the Census and Manpower 
 Administration, U.S. Government Printing Office, Washington, D.C., 1974. 
 
 6 See also. Dr. Richard Wirthlin's discussion of the ecological fallacy (p. 
 81). 
 
 97 
 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 . E 
 
 o 
 
 c £ 
 
 o c: 
 <-> 3 
 
 98 
 
CD 
 (J) 
 
 O 
 
 in 
 
 h- 
 o 
 
 Q 
 
 O 
 DC 
 
 < 
 LL 
 O 
 LU 
 
 _J 
 < 
 > 
 
 LU 
 
 (J 
 
 < 
 DC 
 LU 
 
 > 
 
 < 
 
 LO 
 < 
 
 99 
 
100 
 
THE GRAPHIC PRESENTATION OF STATISTICS 
 
 UA-SMSA 2800-6 
 
 MEDIAN FAMILY INCOME 
 
 & <b & 
 
 
 if) 
 
 
 o 1 - 
 
 •»" 
 
 «s 
 
 % 
 
 < 
 
 n 
 
 85.0 or Over 
 
 65.0 - 84.9 
 
 35 - 64.9 
 
 - 34.9 
 
 
 
 
 
 i 
 
 rr 
 
 
 
 
 
 X 
 
 
 
 
 
 u 
 
 
 
 
 
 □ 
 
 Data Not Available 
 
 MAP 7. 
 FORT WORTH, 
 TEXAS 
 
 Standard Metropolitan 
 Statistical Area 
 
 MEDIAN FAMILY INCOME 
 
 gh the cooperation ol the U S Departmenl ol Commerce, Social and 
 ■nation Bu ;au of the Census: US. Depanment ol 
 
 101 
 
THE GRAPHIC PRESENTATION OF STATISTICS 
 
 UA-SMSA 2800-5 
 
 MEDIAN FAMILY INCOME 
 
 > tf <£ O^ 
 
 
 / 
 
 
 
 
 S5.0 or Over 
 
 
 
 
 
 650 - 84,9 
 
 
 
 
 
 35.0 - 64.9 
 
 
 
 
 
 0.0 - 34.9 
 
 
 
 
 
 □ 
 
 Data Not Available 
 
 MAP 8. 
 
 FORT WORTH, 
 
 TEXAS 
 
 Standard Metropolitan 
 Statistical Area 
 
 PERCENTAGE OF ALL PERSONS 25 YEARS OLD AND 
 OVER WHO ARE HIGH SCHOOL GRADUATES 
 
 J S Department of Commerce. Soci 
 ieu of the Census, US Deparim 
 
 102 
 
THE GRAPHIC PRESENTATION OF STATISTICS 
 
 UA-SMSA 2800-7 
 
 MEDIAN FAMILY INCOME 
 
 £> "> _i a 
 
 ** * # # 
 
 -5 Ss 65 % 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 Data Not Available 
 
 MAP 9. 
 
 FORT WORTH, 
 
 TEXAS 
 
 Standard Metropolitan 
 Statistical Area 
 
 INTERRELATIONSHIP OF FAMILY INCOME 
 AND EDUCATIONAL ATTAINMENT 
 
 Economic Slatis 
 
 103 
 
THE CENSUS FUNCTION 
 
 Daniel B. Levine, Associate Director tor Demographic 
 Fields, U.S. Bureau of the Census. 
 
 "Most people working in this area feel that social indicators provide us better 
 answers than we would have without them, if only as a basis for decisionmaking or, at 
 best, that they influence the overall climate of opinion .... The Census Bureau would 
 like to get the most accurate data it possibly can. On the other hand, the Census Bureau is 
 not '1984.' " 
 
 Daniel Levine 
 
 Mr. Barabba: This morning we're going to discuss a little bit 
 more about the Census Bureau. We have with us Dan Levine, the 
 Associate Director for Demographic Fields. What "Demographic 
 Fields" means, basically, is the area of the Census Bureau that 
 uses as its unit of measurement either a person or a household. 
 
 Mr. Levine: I'm really here under somewhat false pretenses, in 
 that the theme of this series of lectures is social indicators. I think 
 I have a little bit of understanding in this area, but certainly I am 
 far from being an expert in the field. 
 
 Before you pick up your books and papers and decide to 
 leave, I do have a few things that I'd like to share with you today, 
 much of which forms the background for the topic of social 
 indicators. I really would like to give you my understanding of 
 social indicators— a little bit of what I think they are, what they 
 ought to be, what the problems are in using them, what people 
 here and elsewhere are doing in the general area of social 
 indicators, and what the Census Bureau is doing, generally, in 
 trying to help with their development. 
 
 By now you have heard several definitions of social 
 indicators. Let me give you mine. Social indicators, as I 
 understand them, are a time series of social and demographic 
 measures, a series of quick snapshots over time, a series of dial 
 readings which provide social policymakers with an understanding 
 of what's going on. I think the key that describes the term "social 
 indicator" is the idea of a "time series." A snapshot of any one 
 point in time is certainly useful, but by itself it has very little 
 utility because it doesn't tell you whether you're moving up a 
 scale or riding on a flat plane. It's very difficult to call something 
 a social indicator if you only have one observation, or one 
 snapshot. You need to have an objective and to be able to 
 measure whether you are moving toward it. That's the key point 
 that I want to make. 
 
 The second point is in relation to the problems that people 
 have in talking about social "indicators. Notwithstanding the goal 
 that everybody's seeking, no one yet has developed what is 
 desired— a single comprehensive measure that could be described 
 as the social indicator equivalent to the GNP, the latter being 
 generally considered the single composite measure of economic 
 indicators. 
 
 Instead, you have a series of different indicators. Despite all 
 the component pieces, nobody has yet developed what you could 
 call a single comprehensive measure of what's happening to health 
 in our Nation, or what's happening as regards social equality, 
 educational performance or the overall aspiration of our people. 
 In other words, there is no index of the quality of life. 
 
 A number of social scientists have raised useful questions 
 
 about social indicators. Many persons attempting to develop 
 social indicators stress the problem of the supply of social 
 information. But far too little attention is really being paid to 
 how a growing supply of technically improved information could 
 or should be used by citizens, policymakers, and public officials. 
 
 A second problem is that indicators, although of general 
 interest, may not be useful to specific users, each of whom might 
 best be served by a variety of different measures. For example, 
 take the police chief, the precinct captain, and civic groups who 
 are worried about police matters, and political leaders who 
 develop programs and pass legislation. All these people want to 
 appraise the war against crime, yet each may require a different 
 measure. 
 
 A final point is that an indicator tends to be most useful to 
 producers of public goods and services rather than to the 
 consumers of these items. For example, there are indicators that 
 are used by school officials or health planners to tell them the 
 best utilization of health manpower, school manpower, or dollars 
 spent. But, how do you judge the competence of physicians if 
 you happen to be someone who's a patient rather than another 
 doctor? You may have measures of the time it takes a physician 
 to see a number of patients, and how much it costs to train him. 
 But there's a problem as to how social indicators meet the needs 
 of the ultimate user rather than the producer. 
 
 I hasten to add that these reflections are not ideas that I 
 made up, rather, they are contained in a recent report of the 
 Urban Institute. It describes a 2-year program through which 
 they are trying to develop some methodology and determine how 
 useful social indicators can be and how best to structure social 
 indicators so policymakers at all levels can use them. There even 
 are some heretics who believe the development of adequate 
 theoretical underpinnings for social indicators is the most 
 important requirement. They argue that people are zealously 
 developing social indicators but very few people have really 
 stopped to develop a theory of how the indicators interrelate, 
 how they should be used, and what they should be used for. 
 
 After one looks at all these caveats a single point still comes 
 up, namely, that most people working in this area agree social 
 indicators provide us better answers than we would have without 
 them, if only as a basis for decisionmaking or, at best, that they 
 influence the overall climate of opinion. 
 
 Several facts seem certain. There aren't enough data to 
 meet all the needs that people seem to express which is amazing 
 when you contemplate all of the statistics' we have. The fact 
 remains when people start looking for pertinent data regarding a 
 social problem, invariably what they need is missing. A very good 
 example is crime. Crime existed before the earliest law codes. But 
 
 104 
 
THE CENSUS FUNCTION 
 
 there weren't a lot of statistics on crime until it was determined a 
 very important social issue, necessitating the development of 
 public policy and appropriation of public funds. Then, and only 
 then, did someone say, "Where are the statistics which will help 
 us develop the programs we need, statistics which will enable us 
 to measure over time where we're going, where we've been, and 
 whether we are meeting the objective we set for ourselves?" 
 Every scientist will tell you there is unfortunately not enough 
 data around. 
 
 The second point is that social indicators are here to stay. 
 What's going to happen now is that people are going to try to 
 develop bigger and better social indicators and hopefully find the 
 underpinning to the general social theory that I mentioned. 
 
 This leads me to the point that social indicators are not 
 solely the concern of the United States. They are of interest 
 throughout the enttre world. Most of the countries in Europe, 
 both through their own efforts and in conjunction with the 
 Organization for European Economic Cooperation and Develop- 
 ment, have gone a long way toward committing themselves to a 
 continuing program of developing and publishing social indicators 
 of one sort or another. 
 
 The important development here is that these nations 
 started from a very favorable point of attack. They agreed they 
 would try to establish common denominators so that in attempt- 
 ing international comparisons one would have a common frame- 
 work or set of definitions. Hopefully, there also would be a 
 common set of measures so that comparisons could be made as to 
 what was happening in the health field in these different 
 countries. That type of comparative information could serve as a 
 basis for seeing if the programs that they have applied would 
 meet other nations' goals instead of worrying more about 
 conceptual differences to explain away discrepancies in the levels 
 of the social indicators. 
 
 The following table lists the social concerns that five coun- 
 tries-the United States, Great Britain, France, Sweden, and Nor- 
 way—have agreed upon. 
 
 In looking at the table it is rather apparent that basically 
 the same eight topics appear in all of the reports under more or 
 less the same names: Health, safety, education, employment, 
 income, housing, leisure and recreation, and population. You'll 
 notice that some countries have a few more, some countries a few 
 less. These are the social concerns which social scientists from 
 these countries have discussed in concert, the basic concerns that 
 they want to attack in terms of developing social indicators that 
 can help society to understand these problems. 
 
 These categories are interesting in that the countries started 
 off their research individually and through their own efforts came 
 to roughly this grouping. When they got together and decided to 
 discuss and rank-order them, there was a startling consensus. This 
 is particularly impressive when you deal in the international 
 sphere, where each country has its own political and social 
 philosophy and national objectives. This consensus now extends 
 to Japan and a number of the countries in central Europe, 
 including some behind the Iron Curtain where the social 
 philosophy is obviously quite different. There is one area which is 
 left out of the United States' report called "social services " 
 "social welfare," or "social benefits," as you see reported 
 elsewhere. 
 
 You have in front of you the report put out by the Office 
 of Management and Budget, Social Indicators 1973. In talking 
 with the people who worked on it and having participated to a 
 very limited extent myself, I understand the exclusion of social 
 
 services in the U.S. report is a result of a decision to report 
 output measures only. Social welfare and social benefits are 
 viewed as input measures. As you can see, there are differences of 
 opinion as to the best approach! 
 
 Student: How does the United States compare with Sweden and 
 France who have more variables? 
 
 Mr. Levine: As indicated, you'll find these basic eight categories, 
 but you do find in Europe a considerably greater interest in the 
 input measures and in breaking down some of these categories 
 into several component pieces. I think that's the major difference. 
 In a number of those countries, because of the forms of 
 government and because of the statistical systems, their concerns 
 are quite different. In the Scandinavian countries, as you know, 
 they've had social programs such as health insurance for many 
 years, while we're basically just beginning to scratch the surface 
 of a general health insurance program. Where many European 
 countries have long had social programs with far greater develop- 
 ment and participation on the part of their population, plus 
 greater degrees of regulation of the society and the economy, I 
 believe there is a much greater interest on the part of the 
 policymakers to get some measure of the impact those programs 
 are having. The United States has a great deal of interest in social 
 indicators. But it has some national political and social differ- 
 ences as to the role of government. The United States at this 
 point is not quite ready to make a political judgment that it 
 knows what it wants and where it wants to go with that type of 
 measure. That's the basic difference, I think. 
 
 Mr. Barabba: It's a lot easier for a country with a population the 
 size of California to get something going than it is to try to bring 
 in all the cultural and subcultural differences that exist within a 
 country the size of the United States. Also, the amount of change 
 that's taken place within the United States, over any period of 
 time, causes you to come up with different programs. Dr. Jones 
 has shown that our economic growth is the size of most 
 countries. We have to have a continually changing kind of 
 program. That's one major reason why it's very difficult to think 
 about it in a more precise set of terms. 
 
 Mr. Levine: If you look at our political and social development 
 and go back only a short time, you'll find that there wasn't a lot 
 of Government program intervention in the public area. Go back 
 to the 1930's, and there was very little governmental activity. 
 Even in the 1940's you'd only had a little bit, and it was more in 
 the area of economic activity. 
 
 Then you come to the 60's, where you have the vast 
 programs and focusing of attention in this country on poverty, 
 discrimination, and the social measures to improve people's per- 
 ception of the way they live. This called for a very different eval- 
 uation. It wasn't just the same laissez-faire attitude 
 
 The view towards people has changed drastically. It's no 
 longer that people are on the public dole; rather, it's an attempt 
 to assist them to be useful, worthwhile, and participating 
 members of the society. In this sense we have lagged in our 
 development because we're a young country in terms of our 
 history. That's the reason that you see many more concerns in 
 terms of social services among the European nations. Its a 
 temporal thing more than anything else. 
 
 Student: I can agree with that. I grew up in poverty but 
 unfortunately I didn't know it until I was 20 years old. 
 
 105 
 
THE CENSUS FUNCTION 
 
 
 
 
 
 
 c 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 •H 
 
 
 
 c 
 
 
 
 
 
 
 P 
 
 
 
 "• 
 
 
 
 
 
 
 Ci. 
 
 
 
 ■H 
 
 
 
 
 
 
 H 
 
 
 
 P 
 
 
 
 
 
 
 3 
 
 
 
 CO 
 
 
 
 
 
 
 7. 
 
 in 
 
 
 a 
 
 
 
 
 
 
 3 
 
 V 
 
 
 ■H 
 
 
 
 
 
 
 
 
 o 
 
 
 o 
 
 
 
 
 
 
 u 
 
 •H 
 
 > 
 
 
 •H 
 P 
 
 c 
 
 
 
 
 P 
 
 Ti 
 
 Eh 
 
 
 Sh 
 
 
 
 
 c 
 
 
 c 
 
 e 
 
 
 
 
 id 
 
 H 
 
 
 
 
 
 V 
 
 cd 
 
 CO 
 
 
 a. 
 
 p 
 
 
 ■H 
 
 h(l 
 
 H 
 
 
 
 
 
 aj 
 
 a 
 
 P 
 
 C 
 
 >. 
 
 
 
 H 
 
 
 ^ 
 
 .— i 
 
 p 
 
 cd 
 
 •H 
 
 c 
 
 = 
 
 cd 
 
 
 
 cd 
 
 3 
 
 r-H 
 
 u 
 
 7: 
 
 rH 
 
 
 
 
 6 
 
 ■rH 
 
 D, 
 
 cd 
 
 3 
 
 3 
 
 a 
 
 
 
 a 
 
 ■H 
 
 o 
 
 
 
 0> 
 
 w 
 
 
 
 1 
 
 a 
 
 
 
 u 
 
 
 
 (V 
 
 a 
 
 U 
 
 Hi 
 
 M 
 
 CO 
 
 u 
 
 CO 
 
 m 
 
 -j 
 
 -H 
 
 
 
 J3 
 
 o 
 
 
 p 
 
 
 
 en 
 
 
 en 
 
 
 
 
 
 
 P 
 
 
 
 d 
 
 
 p 
 
 en 
 
 — 
 
 
 HMCI^iniDhOOCD 
 
 c/) 
 
 I— 
 or 
 o 
 
 Q_ 
 
 LU 
 
 cn 
 
 o 
 
 < 
 
 Q 
 
 < 
 
 o 
 o 
 CO 
 
 o 
 
 CO 
 
 I- 
 
 LU 
 
 O 
 O 
 
 >• 
 
 
 cr 
 
 
 < 
 
 cd 
 
 ^ 
 
 '0 
 
 - 
 
 S 
 
 'J 
 
 D 
 
 
 CO 
 
 
 — 
 
 
 V 
 
 3 
 
 <H 
 
 hr 
 
 " 
 
 
 
 - 
 - 
 
 rH 
 
 11 
 
 3j 
 
 cd 
 
 o 
 
 
 ■H 
 
 •H 
 
 ~ 
 
 'J 
 
 <H 
 
 c 
 
 '; 
 
 <H 
 
 cd 
 
 co 
 
 O 
 
 
 
 
 P 
 
 P 
 
 en 
 
 
 
 
 
 P 
 
 c. 
 
 
 
 
 P 
 
 
 ■a 
 
 
 
 
 3 
 
 P 
 
 
 
 X 
 
 ■H 
 
 B 
 
 
 CI) 
 
 
 7) 
 
 
 3 
 
 
 3 
 
 
 
 3 
 
 
 
 
 ~ 
 
 3 
 
 
 Sh 
 
 
 3 
 
 
 cd 
 
 
 Sh 
 
 P 
 
 e 
 
 p, 
 
 CD 
 C 
 
 
 
 — 
 
 3 
 
 3 
 
 n 
 
 en 
 
 3 
 
 
 cd 
 
 
 >-3 
 
 
 c 
 
 
 co 
 
 
 
 
 
 
 P 
 
 3 
 
 U 
 
 - 
 
 
 rH 
 
 
 T, 
 
 
 
 
 
 
 
 i-H 
 
 3. 
 
 
 3 
 
 
 
 u 
 
 
 o> 
 
 3 
 
 3 
 
 
 ■H 
 
 u 
 
 CJ 
 
 11 
 
 e 
 
 
 V 
 
 rH 
 
 U 
 
 
 
 is 
 
 
 
 cd 
 
 
 
 P 
 
 3 
 
 ■H 
 
 > 
 
 
 
 hn 
 
 H 
 
 cd 
 
 ■H 
 
 o 
 
 
 
 ■H 
 
 
 rH 
 
 cd 
 
 P 
 
 e 
 
 CJ 
 
 u 
 
 e 
 
 >, 
 
 3 
 
 - 
 
 p 
 
 J3 
 
 H 
 
 P 
 
 n 
 
 3 
 
 
 o 
 
 
 
 J-J 
 
 
 ■H 
 
 
 
 n 
 
 '' 
 
 cd 
 
 p 
 
 cd 
 
 cd 
 
 H 
 
 en 
 
 3 
 
 3 
 
 3 
 
 
 3 
 
 CO 
 
 ^H 
 
 en 
 
 3 
 
 > 
 
 rH 
 
 ■rH 
 
 o 
 
 -H 
 
 ■H 
 
 P, 
 
 Eh 
 
 
 
 - 
 
 o 
 
 3 
 
 P, 
 
 u 
 
 " 
 
 -rH 
 
 cd 
 
 
 
 3 
 
 Eh 
 
 "J 
 
 
 
 p 
 
 CJ 
 
 3 
 
 
 
 
 £ 
 
 CD 
 
 u 
 
 Ih 
 
 
 
 R 
 
 — 
 
 I 1 
 
 J 
 
 a, 
 
 CO 
 
 Ul 
 
 cd 
 
 cd 
 
 P 
 
 cd 
 
 X 
 
 ft 
 
 w 
 
 ft 
 
 « 
 
 M 
 
 rj 
 
 
 
 . 
 
 
 
 
 
 
 
 
 
 . 
 
 
 
 3 
 
 rH 
 
 rH 
 
 
 M 
 
 
 Q 
 
 CO 
 
 -r 
 
 LO 
 
 
 CO 
 
 I> 
 
 X 
 
 0! 
 
 rH 
 
 rH 
 
 
 
 
 
 >. 
 
 3 
 
 
 
 
 
 
 
 
 S 
 
 1) 
 ■0 
 
 
 
 
 
 
 3 
 
 
 
 rH 
 
 
 <P 
 
 
 a 
 
 a 
 
 
 
 
 T) 
 
 u 
 
 
 
 
 
 
 
 
 
 ■H 
 
 
 3 
 
 CJ 
 
 en 
 
 
 
 ■H 
 
 ■H 
 
 P 
 
 
 cd 
 
 i- 
 
 p 
 
 p 
 
 
 P 
 
 P 
 
 n 
 
 
 
 
 o 
 
 en 
 
 
 cd 
 
 cd 
 
 rH 
 
 
 1! 
 
 T) 
 
 ■H 
 
 
 
 
 r-i 
 
 rH 
 
 3 
 
 
 h 
 
 
 
 H 
 
 U 
 
 
 3 
 
 3 
 
 ft 
 
 p 
 
 ■H 
 
 P 
 
 P 
 
 
 
 a 
 
 a 
 
 
 
 3 
 
 H 
 
 cd 
 
 3 
 
 — 
 
 
 r 
 
 "' 
 
 a. 
 
 il 
 
 
 rH 
 
 ': 
 
 3 
 
 
 u. 
 
 ft 
 
 
 
 r_ 
 
 bo 
 
 
 
 U 
 
 cd 
 
 tit 
 
 CD 
 
 rH 
 
 > 
 
 '.! 
 
 •H 
 
 1 
 
 rH 
 
 en 
 
 3 
 
 > 
 
 cd 
 
 ■H 
 
 
 _<: 
 
 X 
 
 '' 
 
 0) 
 
 ■H 
 
 ■H 
 
 p 
 
 P 
 
 C. 
 
 rH 
 
 rH 
 
 a 
 
 hi) 
 
 > 
 
 P 
 
 
 
 C> 
 
 | 
 
 '5 
 
 n 
 
 cd 
 
 rt 
 
 ■H 
 
 ci 
 
 H 
 
 < 
 
 £ 
 
 s 
 
 J 
 
 tS 
 
 J 
 
 < 
 
 
 
 
 
 
 
 
 
 
 
 3 
 
 
 
 
 O 
 
 
 
 
 -H 
 
 P 
 
 
 
 P 
 
 3 
 
 
 en 
 
 P. 
 
 Cl> 
 
 
 3 
 
 f= 
 
 i 
 
 - 
 
 It 
 
 3 
 
 c 
 
 3 
 
 
 
 en 
 
 ■H 
 
 CC! 
 
 P 
 
 3 
 
 3 
 
 
 ■M 
 
 
 
 - 
 
 '/I 
 
 3 
 
 (J 
 
 M 
 
 rH rH O 11 *H bO 
 
 O O -H rH P C £ 
 
 J3 £ P 3 Cd -H P 
 
 1) u cd en cj en i-h 
 
 en en o 'H 3 3 cd 
 
 
 ■n 
 
 ■H 
 
 3 
 
 •H 
 
 
 en 
 
 H 
 
 
 
 n 
 
 P 
 
 
 Cj 
 
 3 
 
 3 
 
 ■H 
 
 3 
 
 
 H 
 
 ■H 
 
 CI) 
 
 p 
 
 £i 
 
 
 p 
 
 rH 
 
 m 
 
 3 
 
 ■H 
 
 
 H 
 
 hQ 
 
 
 X2 
 
 3 
 
 CD 
 
 Eh 
 
 ■H 
 
 rH 
 
 -H 
 
 p 
 
 3 
 
 rt 
 
 rH 
 
 cd 
 
 tn 
 
 en 
 
 3 
 
 a 
 
 ■H 
 
 •H 
 
 P 
 
 ■H 
 
 :j 
 
 en 
 
 3 
 
 CJ 
 
 en 
 
 TI 
 
 > 
 
 rH 
 
 rf 
 
 c 
 
 ■H 
 
 3i 
 
 
 
 3; 
 
 UJ 
 
 CO 
 
 a 
 
 os 
 
 a 
 
 Cd Sh rH O Cd CD 
 
 > CJ H 10 H E 
 
 o u 
 
 CO o 
 
 rHrHCNC»J^inCX)r~ 
 
 — 
 
 Eh rH 
 
 
 
 3 
 
 3 cd 
 
 
 
 3 
 
 P C 
 •rH 
 
 
 
 CD 
 
 ■a en 
 
 en 
 
 
 6 
 
 (3 rH 
 
 n 
 
 
 
 
 1) CD 
 
 Cl 
 
 
 CJ 
 
 a Oh 
 
 •H 
 
 
 3 
 
 M 
 
 > 
 
 
 •— ( 
 
 W XJ 
 
 rH 
 
 
 
 c 
 
 CD 
 
 3 
 
 rH 
 
 rH Cd 
 
 CO 
 
 C 
 
 -H 
 
 3 
 
 rH03OP0PCdCd-rtrHP0 
 
 rHenrH-HCjco.iHenen 
 
 !HCJ-;Cj2rHPCO 
 
 a a 
 s 
 
 CD CD CD 
 J Oh 3: 
 
 > 3 CD 11 C 
 
 C r-J CC a rH 
 
 h n n rf 
 
 p 
 
 CD 
 
 J3 CJ P >> 
 p -H cd O 
 
 <-l ^1 CJ f-1 
 
 P. o 
 
 r3rH 
 
 3 
 
 a. 
 
 
 HNcorfmcoiNco 
 
 a 
 
 3 
 
 (J H CD 
 
 P BJ 
 
 CD P 
 
 c cd 
 
 CD 
 
 CD rH 
 
 X! O 
 
 P >-^ 
 
 o -a 
 
 P CD 
 
 •o 
 
 CD TJ 
 
 1) X) 
 C C 
 O cd 
 
 CD 
 
 en 
 
 CD 
 
 rH 
 
 Sh 
 
 
 Eh 
 
 CD 
 
 ■> 
 
 a 
 
 .3 
 
 CD 
 
 > 
 
 P 
 
 'rH 
 
 
 
 
 CJ 
 
 ■/, 
 
 
 J3 
 
 H 
 
 en 
 
 
 en 
 
 ■rH 
 
 ctl 
 
 3 
 
 
 
 
 p 
 
 cd 
 
 
 ■H 
 
 rH 
 
 rH 
 
 
 rH 
 
 •H 
 
 
 
 
 ■H 
 
 
 CO 
 
 cd 
 
 cd 
 
 
 
 » 
 
 
 
 en 
 
 J= 
 
 
 
 
 ■H 
 
 s 
 
 
 7. 
 
 
 
 
 
 tn 
 
 CJ 
 
 
 ■H 
 
 CJ 
 
 
 
 Ih 
 
 H 
 
 
 
 
 
 >, 
 
 
 P 
 
 
 
 - 
 
 
 0) 
 
 rH 
 
 
 
 p 
 
 cn 
 
 
 H 
 
 7^ 
 
 c 
 cd 
 
 
 p 
 
 a 
 
 
 
 IH 
 
 
 
 3 
 
 a 
 
 
 cd 
 
 3 
 
 >» 
 
 
 
 3 
 
 
 
 
 •H 
 
 
 
 
 en 
 
 -3 
 
 — 
 
 
 .3 
 
 
 
 
 
 ~ 
 
 jd 
 
 
 3 
 
 
 3 
 
 
 O 
 
 x: 
 
 
 
 3 
 
 p 
 
 
 CJ 
 
 — 
 
 a 
 
 
 
 p 
 
 
 
 M 
 
 
 hfi 
 
 
 Cj 
 
 rH 
 
 
 CM 
 
 
 
 
 
 •a 
 
 3 
 
 TJ 
 
 rH 
 
 tit 
 
 
 rH 
 
 en 
 
 >> 
 
 
 
 3 
 
 •H 
 
 3 
 
 01 
 
 3 
 
 
 
 3 
 
 Eh 
 
 
 3 
 
 cd 
 
 P 
 
 3 
 
 > 
 
 w 
 
 
 
 
 •H 
 
 cd 
 
 
 ■H 
 
 
 3 
 
 
 
 
 
 
 s 
 
 cd 
 
 p 
 
 
 CJ 
 
 en 
 
 3 
 
 cn 
 
 CJ 
 
 >. 
 
 
 p 
 
 p 
 
 3 
 
 
 
 
 cn 
 
 C 
 
 p 
 
 
 rH 
 
 
 
 3 
 
 CU 
 
 
 CO 
 
 OJ 
 
 CJ 
 
 a 
 
 >> 
 
 3 
 
 
 
 
 
 
 E 
 
 >> 
 
 
 3 
 
 CJ 
 
 
 
 Eh 
 
 
 
 
 p 
 
 o 
 
 § 
 
 rH 
 
 
 
 rH 
 
 -c 
 
 TJ 
 
 
 
 
 
 
 
 C 
 
 r( 
 
 > 
 
 
 
 •H 
 
 p 
 
 rH 
 
 
 3 
 
 
 
 o 
 
 
 
 •H 
 
 >-< 
 
 rH 
 
 o 
 
 ■r^ 
 
 
 
 
 O 
 
 3 
 
 
 ~ 
 
 P 
 
 
 cd 
 
 CJ 
 
 u 
 
 > 
 
 
 ■H 
 
 en 
 
 H 
 
 ^ 
 
 CJ 
 
 r~i 
 
 Sh 
 
 <: 
 
 Eh 
 
 
 
 B 
 
 P 
 
 en 
 
 cd 
 
 PJ 
 
 H 
 
 cd 
 
 3 
 
 
 CD 
 
 CJ 
 
 
 
 ■H 
 
 •H 
 
 ■H 
 
 
 r~ 
 
 
 P 
 
 X) 
 
 P 
 
 
 X) 
 
 ~ 
 
 
 u 
 
 
 
 A 
 
 3 
 
 rH 
 
 3 
 
 
 cd 
 
 hB XJ 
 
 CO 
 
 
 
 J3 
 
 3 
 
 
 
 3 
 
 
 
 
 
 p 
 
 c 
 
 cd 
 
 i> 
 
 CO 
 
 H 
 
 CO 
 
 s 
 
 O 
 
 BCJ 
 
 ■fi 
 
 cd 
 
 ■rH 
 
 
 a> 
 
 
 
 
 
 
 
 H 
 
 — 
 
 
 rfl 
 P 
 
 in! 
 XI 
 
 o 
 p 
 
 
 rH 
 
 o 
 
 rC 
 
 
 
 
 
 
 
 106 
 
THE CENSUS FUNCTION 
 
 Mr. Levine: What I'd like to do now is to turn to the role of the 
 Bureau in providing the data that become social indicators. The 
 Bureau's job is to produce the data that people need. It's really a 
 data factory of sorts. I don't think the Bureau's job is 
 considerably different, whether we're talking about social indica- 
 tors or economic indicators. One way to describe the Bureau is in 
 terms of a story that one of our colleagues, a mathematical 
 statistician, told me many years ago. He told the story of the 
 three umpires. The very young umpire said, "I call them as I see 
 them," and the more mature umpire said, "Well, I call them as 
 they are," and the grizzly veteran said, "They ain't nothing until I 
 call them." 
 
 In any case, the story indicates what happens without, 
 data— you're flying blind! Until you measure something, you 
 don't know where the devil you are. Probably the classic example 
 of that is a program with which I've been associated a long time, 
 namely, the Current Population Survey. This is the survey that 
 develops the employment and unemployment statistics. It really 
 got its impetus back in the late 1930's. This was the time, as some 
 of you may recall, when we had something which we didn't 
 euphemistically call a "rolling readjustment" or a "mini reces- 
 sion." Rather, we had a depression. On the one hand you had the 
 National Association of Manufacturers saying that there were 3 
 million unemployed and the AFL-CIO that there were 1 3 million 
 unemployed. The estimates ranged from one extreme to the 
 other. Finally a survey method called "sampling" was developed. 
 By chance, it reached a point where it could be used, and we 
 began to obtain a constant repetitive measure of unemployment. 
 Since that time, we've been producing measures of unemploy- 
 ment on a monthly basis. Today you not only have a snapshot at 
 a given point in time, but you have a series or a trend, and this 
 enables you to see whether you're going up, down, or sideways. 
 
 What does the Bureau do in the "social" area? The place to 
 start is our Government's single largest effort, the decennial 
 census of population and housing. 
 
 I don't know if any of you have ever heard it referred to as 
 the largest peacetime mobilization effort. I don't think most of us 
 realize how extensive it is. 
 
 For example, in an incredibly short time span of 6 months, 
 we put together and trained a field staff of approximately 
 185,000 part-time people for the 1970 Census of Population and 
 Housing. It was a very large group, and they were spread over the 
 entire Nation, including the territories and possessions. We opened 
 and operated 400 offices, we printed and mailed some 70 million 
 questionnaires, or more, to 70 million households, just before 
 Census Day, April 1, 1970. By December of 1970, we were able 
 to provide the basic population data to the President in order to 
 meet the constitutional requirement for the apportionment of 
 our Congress. It's really rather amazing it could be done in such a 
 short time. Ultimately, we produced some 2,000 reports and 
 200,000 pages of information, not to mention tapes, microfiche, 
 and microfilm, and we did it at a cost of around $1.00 a person, 
 or somewhere around $220 million for 1970. 
 
 Student: Would you explain what a microfiche is? 
 
 Mr. Levine: That's a type of microform. Instead of being a reel 
 of tape like microfilm, it's a square piece of sheet microfilm 
 which is like a 3- by 5-inch card on which you can put a 
 tremendous amount of data. It's a way to store an immense 
 amount of data in a very small amount of space and, instead of 
 having to roll a reel of tape through a reader, you can just pick 
 
 out the indexed microfiche you want, put it in the reader, and 
 save yourself the time of having to deal with a vast quantity of 
 film or turn pages in a publication. 
 
 Let me tell you a little bit about how we took the census. 
 In 1970 we used three forms. The first was a short form which 
 basically asked for a limited number of population and housing 
 characteristics— that is, a list of people in the household and the 
 age, sex, marital status, and race of each person, and a little 
 information about housing, and that was about it. Then we had 
 two longer forms, which, in addition to collecting the limited 
 population data, also obtained considerably more detail from a 
 20-percent sample of the population. The form to be filled out 
 for each household was assigned on a statistically random basis. 
 We used approximately three-fifths of the households in the 
 country. In advance of the census we developed what we call an 
 address register. We went out and bought commercial mailing lists 
 which gave us addresses in the larger metropolitan areas and some 
 adjacent counties, and then several times before the census we 
 had the Post Office take the address list that we purchased and 
 check it. In fact, they made believe that they were going to go 
 ahead and distribute mail to every one of the households in those 
 areas of the United States for which we gave them address cards. 
 They gave us back a list of the addresses for which we had not 
 provided them an address card, so that we could add the missed 
 addresses to our files. Then we prepared questionnaire mailing 
 pieces for every one of the households, which the Post Office 
 "cased" and later delivered. Each time they gave us any new 
 addresses that they might have missed before, for example, any 
 new addresses which had been opened up, new construction, 
 what have you, since the first time they cased the list. In some 
 areas we went out prior to the census and had our interviewers 
 list addresses where we didn't have a commercial address list and 
 then went through the mail procedure. Respondents were asked 
 to complete the questionnaires and mail them back. 
 
 For the remainder of the country the Post Office delivered 
 an unaddressed short questionnaire to every household. Then we 
 had interviewers go around and knock on doors, list the 
 households directly, collect the short forms and interview people 
 as necessary. In addition the enumerator completed a long form 
 for those households designated for the sample. We did this 
 primarily in rural areas where it's very, very difficult to get 
 mailing addresses, particularly where you're dealing with RFD 
 boxes. Post Office boxes, and houses which may or may not have 
 addresses where the only identification is a general description 
 such as, "Down that road, third house on the right with the blue 
 roof." Even so, when you come back 2 weeks later it doesn't have 
 a blue roof anymore! 
 
 Student: What was the proportion of the form sent out as to 
 size? 
 
 Mr. Levine: Throughout the United States, 80 percent of the 
 population received a short form and 20 percent of the 
 population a long form. That was in order to provide consider- 
 ably more detail through the use of sampling techniques and thus 
 limit the burden to the population as a whole. 
 
 We find that there are a number of bases for deciding how 
 you put questions on the census. If we used every question that 
 people have suggested on the decennial census, every member of 
 the population would be sitting around his household for about 
 2 hours answering a wide variety of questions and at the same 
 time he would be on the telephone calling his Congressman to 
 
 107 
 
THE CENSUS FUNCTION 
 
 complain bitterly about the invasion of his privacy, and using 
 "expletives deleted" to tell him what he thought about the 
 Census Bureau! Therefore our key criteria are cost, time and 
 need. 
 
 In many cases, the reason a 100-percent count is conducted 
 is that the Constitution mandates it; you must count every person 
 in the population in order to apportion the Congress. To 
 determine certain types of characteristics, however, you don't 
 need to interview every person. Sampling theory is well advanced, 
 notwithstanding the fact that there still are skeptics. Used and 
 controlled properly, sampling can provide data for various 
 geographic levels that serve the needs of policymakers. 
 
 The major decision as to the extent of a survey, aside from 
 the fact that you must count everybody for apportionment, is 
 mainly one of determining the level of accuracy required by the 
 uses of the data products. If you need geographic detail down to 
 the block level, you're going to have to ask everybody the same 
 question. If you need detail only for the city of Los Angeles you 
 don't have to ask the full population, you can ask a sample. We 
 try to determine, as best we can, the uses to which the data are 
 going to be put. The needs consequently determine the type of 
 geography and ultimately the required sample size. For example, 
 in order to meet some of the specified requirements, we had 
 certain questions on both the 15-percent and 5-percent census 
 questionnaires so that we could combine them and, in effect, 
 come out with a 20-percent sample. The point I have already 
 made is that if you put every question on the questionnaire, 
 ignoring the possible reaction of the population, the cost would 
 go up prohibitively because the more data you collect the more 
 you have to process, and processing costs are inordinately high. 
 
 If you take a category such as occupation, which is a 
 description by an individual of what he does, you have to have a 
 clerical staff which reads every one of those entries and converts 
 them into a numerical equivalent in order to get it into the 
 computer and tabulate it. If you extend that occupation question 
 to the entire population the cost will go up by a function of ten 
 or twenty and you'll have real chaos. Thus, a major problem is 
 trying to decide whether a question should go on the 5-percent, 
 the 15-percent, or on both questionnaires, or whether it has to be 
 asked of the whole population. What we've tried to do is 
 minimize the 100-percent items asked of everybody to decrease 
 the burden on the population, and yet maximize the amount of 
 data we collect by geographic detail for use in policy planning. 
 
 Student: Are they proportioned on a random basis? 
 
 Mr. Levine: Yes, straight across the board. For example, out of 
 100 addresses, 15 received the 15-percent questionnaire, 5 got 
 the 5-percent and 80 percent received the "short" form. In 
 addition certain questions appeared on both the 5- and 15- 
 percent forms, so in effect, one in five of the population provided 
 answers to these selected questions. 
 
 Student: So you don't know the names of the people who are 
 going to get the questionnaires? 
 
 Mr. Levine: We don't know the names of anyone to whom we're 
 distributing questionnaires. We deliver the forms to an address, 
 for example, to 801 Kearney Avenue, Apt. 204, Los Angeles. 
 
 Student: Do you pretest your questionnaire before you actually 
 put it in the field? 
 
 Mr. Levine: Rather extensively. Prior to 1960, we had not used 
 the mail-out/mail-back approach. In 1960, we went to a modified 
 technique. In some cases, the interviewer who visited the 
 household to complete the short form left a long questionnaire to 
 be mailed in. Since we had decided to use a "mail-out/mail-back" 
 in 1 970, we spent quite a bit of time and effort during the 1 960's 
 testing the mail techniques because they were brand new. We 
 asked whether we could develop an address register and what 
 kind of mail response rate we might get. 
 
 The shift to the use of a mail approach also reflected the 
 fact that we found it difficult to recruit the large numbers of 
 temporary staff required for the task. Also, the social disorienta- 
 tion that took place during the 1960's caused us a great deal of 
 concern so we decided to explore the use of the mails. We 
 assumed, and our testing proved us right, that we would get a 
 very high response rate in a larger part of the country. This meant 
 that we didn't need as large an amount of resources, in terms of 
 followup capability, in those areas. We received over 95 percent 
 back in the upper income population areas, while in other areas 
 where we had expected difficulty we received a much lower 
 response by mail. Then we were able to use fewer interviewers 
 than expected by focusing our resources on those areas where we 
 knew or anticipated we would have difficulties. We ran at least 
 10 pretests, culminating in 3 rather large dress rehearsals. A 
 pretest is where we go out and experiment in a test setting, where 
 we apply our concepts or techniques to a limited part of the 
 population. A full-dress rehearsal is where we simulate exactly 
 what we are going to do in the census. We did a full-dress 
 rehearsal census in Dane County, Wise; Trenton, N.J.; and 
 Chesterfield and Sumter Counties, S.C. These were the bases on 
 which over the decade we developed the techniques. Our overall 
 response rate in 1970 was about 87 percent for the mail 
 approach, but, as I say, it varied considerably; from a rate of 40 
 to 50 percent in some areas to virtually 100 percent in others. 
 
 We plan in the decade of the 1970's to work towards 
 improving and expanding the mail technique. In 1980, we hope 
 to enumerate some 90 percent of the Nation by the mail 
 approach. 
 
 Student: How do you handle the language diversity in this 
 country? 
 
 Mr. Levine: According to our critics, poorly. Language diversity 
 is one of the major problems that we deal with in a census. We 
 can handle it much more simply in a survey, as I'll indicate later. 
 The way we dealt with it in 1970 was not the way we hope to 
 deal with it in 1980 and certainly not the way we dealt with it in 
 1960. 
 
 The first point, of course, is that in hiring 185,000 
 interviewers and setting up 400 offices, we try to select people 
 who are indigenous to the area in which they're going to work. If 
 we're going to be distributing questionnaires in an area containing 
 Spanish-speaking population, or an area which is Puerto Rican, 
 where they may speak a different type of Spanish than they do in 
 Los Angeles, we try to find people who are interested in working 
 for us who are indigenous to the area and thus are likely to have 
 the language skills required. Second, we make the population 
 aware of the decennial census through a public relations program. 
 As part of the extensive publicity program we provide addresses 
 and telephone numbers of Census assistance centers to be used if 
 anyone has any problems in filling out the questionnaire. 
 
 Further, and this is of extreme importance since there is a 
 
 108 
 
THE CENSUS FUNCTION 
 
 misunderstanding of how we took the 1970 census, we didn't just 
 take what we got back by mail and say, "Thank you very much, 
 the mail returns represent the population of the United States." 
 I'm afraid that isn't what we or Congress would consider a good 
 basis for apportioning that body. Instead, if the questionnaire 
 was not returned, whether because of language difficulties or any 
 other reason, we followed up either by telephone or by sending 
 an interviewer to the household to collect the information, if at 
 all possible, by direct interview. In those cases where the 
 interviewer found that there was a language problem which he or 
 she couldn't handle directly, she went back and got someone who 
 was capable of handling the situation. I would hasten to add that 
 in many households where there is a potential language prob- 
 lem, there are some family members who are generally able 
 to converse in English. It's the older population in particu- 
 lar that has difficulty with language. In many such house- 
 holds, we generally find a younger family member who is 
 able to speak English and can translate from Spanish or 
 Portuguese or Chinese into English. Finally, where we rec- 
 ognized a specific need we hired people with the language 
 skills. In San Francisco, for example, we had a number of 
 interviewers who spoke Chinese or Japanese. Although we did not 
 have a Chinese language questionnaire in 1970, we did provide 
 instruction forms in those languages. In our planning for the 1980 
 census, we are exploring the possibilities of providing the 
 questionnaires in languages other than English. 
 
 We've touched on some of the difficulties faced in the 1970 
 census, namely, the problems of language and attempting to 
 lessen the respondent reporting burden. Prior to the 1970 census 
 there was a fairly extensive campaign by certain groups objecting 
 to the invasion of individual privacy. These groups wanted to 
 know why we asked the individual's name and income, and they 
 objected to many of our questions in general. 
 
 There were some last minute changes made on the 1970 
 questionnaire, resulting from two features. One is that there are 
 rapid changes in the world itself and particularly so in the late 
 1960's. The problems in Watts, Harlem, and Washington, D.C., 
 caught us quite by surprise and we wondered what impact they 
 would have on our ability to take the census. Since it takes a 
 good 6 or 7 years to develop a census, it's difficult if not 
 impossible, to anticipate all of the potential problems. In fact, 
 we're already working full steam on the 1980 census and will 
 begin testing this coming year. 
 
 The second problem was the passage of civil rights 
 legislation and equal rights amendments, which caused many 
 parts of the society to become more aware of the needs and the 
 importance of census data and their particular concerns as to 
 their own self image. 
 
 We, for example, had really never developed an adequate 
 program to measure the participation of the Spanish-origin 
 population in the life of the United States. We provided selective 
 measures, such as Spanish language, citizenship, place of birth, 
 and surname, but we didn't have a uniform measure such as 
 Spanish origin. Reflecting the demands for such information, in 
 May 1969, very late in the census planning, we added a question 
 on Spanish origin in the 1970 questionnaire. This forms the basis 
 for much of the information which we now have on the 
 Spanish-origin population. 
 
 Another unanticipated problem was that of geography, 
 which isn't to say that we hadn't been concerned about 
 geography. Obviously, when we said how many people were living 
 in Los Angeles we wanted that count to be as accurate as 
 
 possible. The attention focused on geography post-1970 is simply 
 far and away much more extreme and intense than in the past. 
 This new concern appears to have surfaced for several reasons. 
 
 As we went to the mail technique we had to generate an 
 entirely new system of putting geographic identification on the 
 1970 census. We developed something called address coding 
 guides where we tried to create lists of streets and address ranges 
 which fell within geographic boundaries. This involved a lot of 
 participation on the local community side and there were real 
 problems because it was the first attempt. I think many people in 
 both the community and the Bureau underestimated the diffi- 
 culty and complexity of really knowing where the boundary lines 
 were. It turned out that, in some cases, the city planners 
 themselves weren't quite sure where the markers were. The 
 marker was somewhere out in Farmer Jones' field, but you 
 couldn't find it. Prior to this there really hadn't been much social 
 legislation tied to economic or geographic boundaries, so there 
 hadn't been much concern. Suddenly, States and local communi- 
 ties were faced with the concerns of "one man, one vote." When 
 you started dealing with "one man, one vote," people became 
 very concerned as to whether they had the right count of 
 population, not only within Los Angeles, but within election 
 district 36 versus district 42, in the city's south side versus the 
 north side. We found that we had made such errors in geography. 
 
 If "one man, one vote" wasn't enough of a problem, Con- 
 gress passed the State and Local Fiscal Assistance Act of 1972, 
 otherwise known as general revenue sharing, and suddenly every 
 head became an amount of money and we started getting letters 
 by the thousands telling us, for example, that we'd counted the 
 population of a place as 187 when it should have been 212. A 
 major problem we face in 1980 is ensuring that we solve our 
 geographic problems so that we can minimize, if not eliminate, 
 this source of error. 
 
 The legislation has helped us to some extent since mayors, 
 local officials, and community activists are aware of the impor- 
 tance of population, they're also aware of geographic boundaries. 
 Each State has established an office which is responsible for 
 maintaining up-to-date records of the current, official boundaries 
 of each general purpose government within the State. Every 
 annexation and every incorporation has to be reported to that 
 particular source. We feel that this development will be of great 
 assistance and make it possible for us to do a much better job in 
 1980. 
 
 Student: That's within this State? 
 
 Mr. Levine: That's within each State. That function is required 
 of the State under the administrative regulations developed in 
 connection with this legislation. 
 
 Student: Is there any special reason why you have to have the 
 name of the individuals to whom you send the questionnaires? 
 Don't you think you'd get a more honest reply and you wouldn't 
 have this problem with confidentiality if individual names weren't 
 involved? 
 
 Mr. Levine: From my point of view the answer is both yes and 
 no. The first question is why do we need the names. It's very sim- 
 ple to look at single family houses in the suburbs and know that 
 if we mail a questionnaire to 801 Kearney Avenue and don't get 
 the form back, we can locate 801 Kearney Avenue, knock on the 
 door and get that questionnaire. However, when you go into the 
 
 109 
 
THE CENSUS FUNCTION 
 
 central city or some of the older areas of a city, there are large 
 houses which have been converted into boarding houses, apart- 
 ment houses, or rooming houses— what we call disorganized 
 housing units. Inside, we find lots of doors without any 
 identification, and we don't know whether they're A, B, C, D, 1, 
 2, 3, 4, and what they were yesterday may not be what they are 
 today. So, without names, we're in trouble! 
 
 Let's assume that I send out 10 questionnaires and I get 4 
 back. We have to send an interviewer out to get the remaining six, 
 and that interviewer will not have the faintest idea which four I 
 got back and which six she has to find. Suppose she goes i n 
 there and doesn't find 10 addresses, or 10 doors, but, instead 12. 
 Then there really is a problem. 
 
 Remember, I also said that we use three forms in the census. 
 We use two scientifically selected sample forms. If the delivery of 
 these forms is mixed up and is not as specified, we may have a 
 biased sample. For example, if the mailman were to have sample 
 forms for apartments 1 and 16 and leaves them in error with 
 apartments 1 and 2, the sample would no longer be controlled. It 
 could be a biased sample. The conclusions that a policy planner 
 drew from that sample's data would be suspect, yet no one would 
 know it. One of the major reasons we need names is to ensure 
 we've counted every unit, in every structure, in every place in the 
 country, and collected the appropriate form for the correct 
 household. 
 
 Secondly, when we get a questionnaire back in the mail, it 
 may have missing information. This may result from a poor 
 questionnaire design, language problems, or other considerations. 
 An interviewer has to call back and clarify the error, inconsis- 
 tency or omission, and it's very difficult to refer to someone in a 
 household without knowing to whom you're referring. Therefore, 
 we find the name is a very useful item which allows us to say, 
 "Mrs. Jones, this is the Bureau of the Census calling. Your 
 husband, Mr. Jones, in filling out his information, or your son, 
 Timothy, in filling out his information, did not answer in a way 
 that we can understand." 
 
 Third, the Census Bureau conducts what we call an age 
 search activity. Age search is an operation which provides people, 
 who do not have proof of birth, with legally accepted identifica- 
 tion which enables them to get passports, veteran's pensions, 
 Social Security, disability insurance, survivors's benefits, or 
 simply to show (a) that they exist, (b) that they are who they say 
 they are, and (c) that they indeed have the relationship to an 
 individual that they say they do. If a person writes to the Census 
 Bureau and says that she lived at 801 Kearney Avenue at the time 
 of the 1970 census and is applying for Social Security under her 
 husband's benefits, the Census Bureau will, for that person alone, 
 not for anybody else, look up the material for 801 Kearney. We'll 
 see if she's listed there, and certify to the Social Security 
 Administration that on April 1, 1970, she was living at 801 
 Kearney and was, indeed, the wife of Mr. Thomas Jones. 
 
 The name, incidentally, is never used in any of the 
 processing activities beyond the collection process. Once we take 
 the questionnaire and put it into a mechanical form for 
 processing, the name is never attached to the coded data. We 
 store the names, for Age Search purposes, at Pittsburg, Kans., 
 but in the millions of computer data tape reels at the Bureau, 
 there isn't a single name. We use a serial number type of 
 relationship to insure confidentiality and because we're simply 
 not interested in names except as I have described. 
 
 Let me go one step further and talk about another reason 
 we do use names. It's not the most important and, perhaps, if the 
 
 issue really came down to where we couldn't do the census 
 another way except to give up names, we might have to. My 
 reference here is to post-censal studies that we're often called 
 upon to conduct. 
 
 Let me give you two examples that illustrate the impor- 
 tance of having names. The National Science Foundation is very 
 interested in knowing what's happening with regard to the supply 
 and training of engineers and scientists in the United States. This 
 happens to be an area of great concern. To develop a new roster 
 of scientists and engineers would require an inordinate cost 
 because they're a rare phenomenon in the population as a whole. 
 You could get lists of the members of professional societies, but 
 not all scientists and engineers belong. Not all people who are 
 working as engineers were trained as engineers so you can't 
 always go back to university lists. 
 
 When we are faced with such a problem, we search through 
 our tapes, select those people who's occupations fit the require- 
 ments, look up their names, and mail them questionnaires. In this 
 case, you're not just mailing to an address. You're mailing to 
 Dr. Jones who is an aeronautical engineer or Mr. Henry Smith 
 who is a physicist. 
 
 Back in the mid-1960's we were asked to participate in an 
 international health survey. There was a concern among the 
 people at the National Institutes of Health that there might 
 possibly be a considerable difference between native born 
 Americans, in terms of the incidence of heart disease and cancer, 
 and some immigrant groups. 
 
 What we did was to take two groups. Norwegians have a 
 relatively high incidence of cancer. We went through the 1960 
 census files and identified people born in Norway and subse- 
 quently interviewed them; without the names we'd never have 
 found them. 
 
 We also took a group of people who had migrated here 
 from the United Kingdom. The British have a relatively high 
 incidence of coronary disease. We also interviewed this group to 
 find out their health status, their types of health problems, and 
 whether they had any siblings living in the United Kingdom or 
 Norway. We arranged with the Norwegian and United Kingdom 
 groups participating in the study to interview these siblings. We 
 also interviewed a sample of the indigenous population in 
 Norway, the United Kingdom and the United States and tried to 
 compare the health morbidity situation. It was an excellent study 
 that provided a great deal of data to the heart and cancer people. 
 Yet without names you couldn't conduct these types of studies. 
 Whether that's a cost benefit to collecting names is something 
 which society must decide. 
 
 Let me make a point. Notwithstanding the fact that the 
 census is mandatory by law and we have the legal right to 
 prosecute someone, we still rely on the voluntary cooperation of 
 the population because we have no way of knowing we've missed 
 somebody if they aren't listed on a questionnaire. If someone 
 returns a questionnaire from household "A" and tells us there are 
 three people, we can't know if there actually were four. The 
 point I was trying to make is the fact that from the vast bulk of 
 the population, well over 95 percent, we receive excellent 
 cooperation. There is no complaint about putting the name 
 down. 
 
 Student: You were talking about most people's voluntary com- 
 pliance. What about where the person doesn't want to answer, to 
 put his name on a form, although he fills out the form completely 
 and just omits his name? Do you go after somebody like that? 
 
 110 
 
THE CENSUS FUNCTION 
 
 Mr. Levine: No. I would say that we wouldn't bother anybody if 
 the form were filled out completely. But the great majority of 
 peopie do put their names down. Only a very small proportion 
 don't. People who don't want to cooperate just don't return the 
 form. 
 
 Legally, the Census Bureau has the right to say, you must 
 fill out everything on the form. As a rule, however, we don't 
 bother the ordinary citizen if most of the form is complete. We 
 may try to get the information by calling back, although we 
 wouldn't call back solely to obtain a name. Now, if the address 
 weren't clear, the interviewer might have to take all of the 
 questionnaires for that particular place along when he or she went 
 back. In other words, if we take the questionnaires along, we can 
 avoid bothering people who have already filled out the forms. 
 
 Student: You have stated that only a few people appear to be 
 really concerned about putting their names on census forms. How 
 does the right of privacy and feelings about privacy here in the 
 United States compare with feelings in some other countries? Do 
 you have more of a problem than other countries, or less, and are 
 the attitudes related to people's feelings about the government? 
 
 Mr. Levine: I think it varies widely. For example, let's take the 
 Scandinavian countries, and the study I referred to a moment 
 ago. As you would expect, most of the Norwegians in our sample 
 came to the United States many years ago, and were well up in 
 their sixties by the time we attempted to contact and interview 
 them. The Norwegian-dwelling siblings were people who were also 
 well up in their years, and, in many cases, surprisingly, they had 
 not been in contact with the American siblings for a long time. 
 Now, most of the Scandinavian countries maintain what are 
 called population registers. It's a very different philosophy than 
 what we have in the United States. Each jurisdiction has a list 
 which includes every person and when he was born, and every 
 time one moves they know exactly where. You can see the 
 concept of privacy there is quite different in terms of having 
 names and addresses available. Even so, they do have problems 
 with privacy in terms of what data are made available from a 
 census and administrative records. But the population registers 
 are open. Anybody can use them. 
 
 In the United Kingdom they have an electoral register, 
 representing every person in the country over 18 years of age, 
 which is made up before each election, checked each year, and 
 published. In Canada they prepare an electoral register before 
 each election and, again, these are published rosters of everybody 
 of voting age. In the United States the reaction would be quite 
 different if we were to publish lists of where everyone lived. So 
 you have different views of privacy and confidentiality. 
 
 The United Kingdom does not collect income information 
 in their censuses, since it is a tremendously controversial item. 
 They did a small sample survey on income and they had hundreds 
 of complaints. 
 
 In some countries, on census day no one's allowed on the 
 street. It's not a legal holiday, you just have to stay in your home 
 until you're counted. 
 
 Student: Do you have any way of knowing from the census 
 information who are illegal aliens? 
 
 Mr. Levine: No, there's no way. 
 
 Student: Do you allow for that? 
 
 Mr. Levine: The census covers everybody in the United States on 
 April 1, 1970. The census contains a question on place of birth 
 and citizenship, but that doesn't tell you whether the respondents 
 are illegal aliens or not, as there are a lot of visitors in the United 
 States who are counted in the census. 
 
 The question of illegal aliens presents another type of 
 problem. If illegal aliens don't want to be counted we don't know 
 how to count them. There's an argument in the Bureau as to 
 whether we do, in fact, get them in the population count. Some 
 people say illegal aliens would never report themselves; other 
 people say that when you go to a household containing only 
 illegal aliens, and somebody answers the door, it's very difficult 
 not to get the count of people there. They may tell you they 
 were born in the USA, but they're illegal aliens and you never 
 know it. You have two possibilities. One is that you get them but 
 you don't know they're illegal, or if they're living in a household 
 where there are six people, you may only have five reported. 
 That's the other problem. There's no way we know of counting 
 illegal aliens. 
 
 Student: Where does the Bureau of Immigration get their 
 estimate? 
 
 Mr. Levine: They get it the same place that I suggested that the 
 CIO, the NAM and other people get their estimates. There are a 
 lot of estimates around. The Bureau of the Census said that there 
 were 10 million people of Spanish origin. Spanish-origin leaders 
 tell us there are 12 or 13 million. We ask them where they get 
 their number and they say, "We know there are 13 million so, 
 therefore, you undercounted by 30 percent." And we respond 
 that we counted 10 million, and how do they know there are 13 
 million? They just know. 
 
 One of the things we do know is that we will never obtain 
 perfection in a census. I don't care if we made everybody stay at 
 home or anything else you wanted. That's the other point I 
 wanted to make about privacy and confidentiality. We're a 
 country that has certain traditions. There's a limit to how far one 
 can go in coercing the population, and we don't want to coerce 
 the population. We go out and try to suggest to people that there 
 are valid reasons for participating in this activity. There are social 
 benefits to be derived from being counted, from having accurate 
 statistics developed. 
 
 It's hard, and we're struggling with it, but from various 
 techniques we've developed we are aware that we undercount the 
 population. The Bureau estimated in 1972 that we undercounted 
 the population in 1970 by 2.5 percent or approximately 5.3 
 million people. Now, 2.5 percent is something that sounds great if 
 you accept the view that we should stop looking at the hole and 
 instead look at the doughnut, where we got 97.5 percent of the 
 population. That's great! We can break our arms patting ourselves 
 on the back, but our critics don't allow us that luxury because 
 there's revenue sharing, and social programs which go to minority 
 groups. The hard fact remains that our analysis, though it is an 
 estimate, suggests that we missed about 1.9 percent of the white 
 population and about 7.7 percent of the black population. We 
 want to do a better job in 1980 and a better job in 1990 and even 
 better in the year 2000. But, I'm sincere when I say I don't think 
 we'll reach perfection. 
 
 At the other extreme, we may reach the point where we 
 overcount the population and this is a major concern, as well. It's 
 almost as bad as undercounting. If people become very much 
 aware of the fact that revenue sharing is apportioned to some 
 
 111 
 
THE CENSUS FUNCTION 
 
 extent on population, there could be a tendency for an organized 
 campaign to develop in certain communities, to add an extra 
 child to the questionnaire. If we get a questionnaire back and it 
 has two children on it instead of one, or three children instead of 
 two, there's no way that we're going to be able to tell you that 
 it's wrong, except through an ex post facto analysis which will 
 say that it looks strange, but for once we have apparently 
 overcounted a specific age-sex-race group. 
 
 We're currently spending an inordinate amount of time 
 trying to develop the techniques we will use in 1980 to minimize 
 the possibility of missing population. Actually, we began this 
 development in 1970. I think we recognized there were areas of 
 the country, like inner cities where we were going to have 
 problems and therefore we put much greater resources in those 
 places in the way of interviewing staff and offices. We had a larger 
 public relations campaign directed toward these areas. We had a 
 community services staff which talked to community leaders to 
 explain what the census was about, in order to get their support, 
 and to get those leaders to go on the radio stations, mount the 
 pulpits, and what have you, to explain the importance of being 
 counted in the census. 
 
 We plan to extend that tremendously in 1980. We plan to 
 deal with all levels of community leaders and their representatives 
 much earlier in the decade of the 1970's. We have already started 
 discussions with Spanish and black groups in order to try to 
 explain why the census is important to them, what we can do for 
 them, and, hopefully, to have them tell us the problems they 
 perceive in our credibility. We want to discover how we should 
 reach their population groups, what's wrong with the question- 
 naire, what types of data they want and, in turn, how they can 
 use the data, when we will produce it, and how we will produce it. 
 
 I've been asked a question as to whether the current 
 Administration, or any Administration, has an impact on what we 
 do at the Census Bureau. I would say, unequivocally, that the 
 methodology is a function of the technicians deciding what's best 
 and demonstrating to policymakers that such and such is the 
 technique to use and this is the way it will go. There's no 
 question about the fact that Congress plays a very important role 
 in this. Every one of the agencies in Washington has to report to 
 Congress, through an oversight committee that looks at it from its 
 operational standpoint and in terms of double jeopardy through 
 the budget process. 
 
 Theoretically, we could come up with a technique that 
 showed the best way to do the census was to make everybody 
 stay at home, but I'm sure that would be a political decision. The 
 Congress would probably replace the technicians who advocate 
 such a course. Or we could decide that the census should cost, in 
 order to do the job right, $800 million and Congress might take a 
 jaundiced look and judge whether the social benefit of an $800 
 million expenditure outweighed the economic problem that such 
 an expenditure would pose. 
 
 Congress is more concerned with privacy, confidentiality, 
 and the individual reporting burden, than it is with the actual 
 questionnaire content. 
 
 Mr. Barabba: One has to keep in mind that there is an extensive 
 and time-consuming process involved in developing the forms and 
 materials for 1980. When the technical staff is ready to submit 
 the materials for congressional review, the Congress is faced with 
 reviewing, in a very short period of time, a volume of evidence 
 developed by a systematic and professional approach to what's 
 needed. Since the census is such a general statistical instrument, 
 
 112 
 
 the other Federal agencies who use it also make input to our 
 questions. If it is decided to change or modify an item, ideally, 
 we would want to have the concurrence of those other agencies. 
 It's not impossible; a question can be changed or modified as then 
 Commerce Secretary Stans did with our question on "sharing 
 showers,"— not a political decision, but a reaction to and a 
 recognition of the sensitivity relative to the need. 
 
 Mr. Levine: The content of the census is not developed in a 
 vacuum. As I indicated, we -alk all over the country with 
 community leaders, policymakers, mayors, and planning commis- 
 sions. In addition, we invite all the Federal agencies to comment, 
 in order to determine their needs for data. Federal agencies are, 
 after all, a reflection of their own Secretaries, and the Secretaries 
 are political. Late in 1967, President Johnson's Cabinet reviewed 
 the plans for the 1970 census, and there were very serious 
 questions raised by the Secretaries of Labor and HEW, because 
 they felt that the programs of the Administration required data 
 which would not be collected in the 1970 census. I think that's a 
 valid question to raise. I don't call that a political decision. But I 
 think you're right, it does reflect the political philosophy or the 
 social philosophy of an administration. They were particularly 
 concerned about poverty at that point and they didn't feel there 
 were enough questions on the census relating to the causes of 
 poverty. 
 
 As a result of the White House's concerns, we went forward 
 with a supplemental request to Congress for some $10 million to 
 conduct post-censal surveys, called the Census Employment 
 Surveys, which were carried out in the ghetto areas of 51 cities 
 throughout the United States. We chose 60 areas in 51 cities and 
 conducted detailed surveys of the reasons that people weren't 
 looking for work, the reasons that they were working less than 
 full time, and asked questions about all of the factors which were 
 felt to be germane to the problem of why the economy wasn't 
 supporting these people adequately. That type of activity 
 certainly reflects the social needs of the moment. 
 
 What I've handed out is the 1970 census questionnaire. (See 
 appendix.) These questions can be grouped into about five 
 components. You begin with the basic demographic items, 
 followed by migration, ethnicity, social and economic aspects. As 
 I said, this is a somewhat arbitrary grouping. 
 
 The demographic questions include age and sex, and you 
 can see those right inside the front cover. For migration we 
 included State or country of birth, place of residence 5 years ago, 
 and year moved into present house. 
 
 Under ethnicity we have color or race, Spanish origin or 
 descent, country of birth, country of birth of parents, citizenship, 
 mother tongue, and year of immigration. 
 
 Social questions are those relating to family or household 
 relationship, marital status, years of marriage, number of chil- 
 dren, years of school completed, school or college enrollment, 
 vocational training, veteran status, and disability. 
 
 For the economic side, we collected information on 
 employment status, hours worked last week, weeks worked last 
 year, the last year in which somebody worked, occupation, 
 industry, class of work or activity 5 years ago, income last year, 
 place of work, and means of transportation to work. 
 
 Student: What is the marital status of a common law relation- 
 ship? 
 
 Mr. Levine: It depends upon how it is reported. If reported as 
 
THE CENSUS FUNCTION 
 
 "never married," then it is "never married," because we have no 
 way of checking, as the designation is up to the individual. Again, 
 the Census Bureau would like to get the most accurate data it 
 possibly can. On the other hand, the Census Bureau is not 
 "1984." 
 
 We rely on the individual and that's one of the reasons we 
 went to the mail technique. We no longer have, in the vast 
 majority of cases, the interaction of an interviewer and the 
 respondent. We felt that when an interviewer knocked on a door 
 there might be some intimidation. Having someone standing 
 there, residents might not want to admit they are in a common 
 law marriage, that a child is illegitimate and they are not married. 
 Also, having an interviewer there implies communication and you 
 might misunderstand the question asked or you might be 
 offended by the intonation in the voice. Mail allows you, within 
 the privacy of your own home, to answer any way you want to. 
 
 We indicate that we want the most accurate information 
 but if you don't consider yourself in a common law marriage, and 
 the key word is "marriage," and you put down single, then our 
 statistics will include you as "single." You and your "roommate" 
 wind up as two single individuals living in one household. We 
 suspect that people of different sex living together in the decades 
 prior to 1970 were reluctant to report themselves as "not 
 married," that they were less concerned in 1970, and may be even 
 less concerned in 1980. We had a much larger number report in 
 1970 than in 1960, that they were unmarried partners of 
 different sexes. In 1980, we think that more people will be 
 willing, because of the increased social acceptability of this living 
 arrangement, to report themselves as co-partners. 
 
 Student: How do you handle head of household if both of them 
 are working? 
 
 Mr. Levine: Working is not the determinant of "head of 
 household." 
 
 Student: Can you have two heads of household? 
 
 Mr. Levine: No, at the present time, you only can have one. In a 
 general sense, our concepts reflect the social acceptance of certain 
 types of arrangements. Up to 1970, the "head" was nominally 
 the married male member of the family. We arc grappling with 
 the problem of what we're going to do in 1980. We really don't 
 care who's the head. What we really need to determine family 
 relationship is to learn every member of that household's 
 relationship to the same individual, so we can put them all 
 together. 
 
 We have done studies of what they call kinship and it gets 
 very difficult in trying to decide, socially or through social 
 theory, who is the head of a household. Is it money? It's not 
 necessarily the main earner in the household. It could be a 
 household which, because of its social orientation, lists the senior 
 member as the head. It can be the wife; it can be any member. 
 Incidentally, again, it's a very small proportion of the population 
 that has raised this question, but it's one we're concerned about 
 nevertheless. 
 
 obtain family and household relationships by determining how 
 each person is tied to the key individual. We don't, at this point, 
 feel we can just say, "put anybody down on that first line." For 
 example, suppose someone enters a child 15 years of age, on 
 the first line. The consistency of the relationship item across 
 families, particularly when examined by characteristics such as 
 income or occupation, would be most suspect, if not totally 
 useless. 
 
 Student: You certainly have to be very careful about this 
 semantically. 
 
 Mr. Levine: Semantics is a real problem. As you look at this 
 questionnaire you can see that it is positive. When you take this 
 and put it in an envelope and include a return envelope, you have 
 a problem in how large you can make it and how much 
 instruction material you can put on it. We've enclosed an 
 instruction sheet but we're not sure that's the best way. Now 
 we're talking about the problem of what we do in 1980 about 
 Spanish language. 
 
 We really have some very serious problems: we have to get 
 a device which will mechanically enable us to handle a paper size 
 which the Post Office will accept, which will go into not only the 
 slot in the mail box but also go into the apartment houses' little 
 boxes, without being crushed so badly that the creases in the 
 paper become marks when we read it through our film optical 
 sensing device. 
 
 Student: What stage of preparation is the Bureau in for the 1980 
 census? How far along are you with testing and developing the 
 questionnaire? 
 
 Mr. Levine: Let me make a key point. Remember, we said that 
 one of the main uses of statistics, whether social or economic, is 
 in trend analysis, and you're pretty well locked into an awful lot 
 of this subject matter. We must, by fiat, count every person in the 
 population. We have to get their age, race, sex, marital status. We 
 want to know their incomes and their work experiences because 
 these are key factors. We want to know fertility. So there's really 
 very little flexibility. What we see, going towards 1980, is 
 probably an expansion of post-censal surveys to attempt to focus 
 in on key subtopics rather than expanding the questionnaire to 
 the point where everybody in the population has to sit down for 
 an hour and a half. 
 
 Secondly, many of the investigations that you want to 
 make are considerably more complex in terms of their organiza- 
 tion and questioning and require a highly skilled interviewing 
 staff. You can only do that if you have a much smaller 
 interviewing staff and use sampling, where you can really devote 
 more of your resources to the quality control and the control of 
 response error, the latter being a key factor. 
 
 Mr. Barabba: Why don't you talk a minute or two of Bureau 
 conversations about questions related to housing, and the 
 trade-off between having the same questions over time and a 
 better measure of housing quality, which may require more or 
 different questions. 
 
 Student: Is it really necessary to designate someone as head? 
 
 Mr. Levine: Yes and no. What is necessary is to designate some 
 "key" member of the household so, as I have stated, we can 
 
 Mr. Levine: You'll note that the 1970 forms contained a series of 
 housing questions. In recent years, serious questions have been 
 raised concerning the adequacy of housing in the Nation and 
 every administration has tried to address this issue since the early 
 
 113 
 
THE CENSUS FUNCTION 
 
 1930's. One of the major problems we have is establishing a viable 
 definition of "good, quality housing," or just "quality of 
 housing," 
 
 As I said, we want trend analysis, but we don't want trend 
 analysis of measures which are no longer applicable or meaning- 
 ful. As the population changes, for example, from agriculture 
 (which was a major factor of the population 50 years ago) to an 
 industrial society which requires many different questions, we 
 have to recognize at some point in time that it is necessary to 
 break continuity, develop new measures, improve concepts, and, 
 if necessary, actually break the series completely and start over 
 again. 
 
 What we are doing in the housing field is investing a 
 tremendous amount of resources in trying to get agreement on 
 such matters, not only among ourselves but among the agencies 
 most concerned, such as Housing and Urban Development and 
 Social Security. Housing quality is one of the major factors in 
 determining levels of living. What used to be "housing quality" 20 
 or 30 years ago is no longer applicable, when you look at what 
 now constitutes an adequate support level. 
 
 For example, to have suggested that someone on welfare 
 should have had a television set, would have resulted in quite a 
 number of raised eyebrows. Now, television is considered a means 
 of communication and a necessity in many places. Therefore, one 
 may qualify for benefits even though a certain amount of income 
 is going into television or certain other amenities. 
 
 In the past we measured housing quality by whether the 
 house was dilapidated or not, whether it had plumbing and 
 electricity. Today, most houses in the United States have 
 electricity. The concern about quality of housing has now shifted 
 to the environment around the household. Is there transporta- 
 tion? Now this becomes a major factor. People can't get to and 
 from work without transportation. What kind of shopping is 
 available? What kind of neighborhood? Is there fear about crime 
 in the neighborhood? 
 
 Today, we're looking at measures of quality of housing 
 which extend beyond the four walls within which people live to 
 the environment around where the people live. Do you have a 
 railroad track in front of the house? People never considered that 
 a major problem before. Now some people do. What's the 
 availability of schools? Where are the buses? So we're trying to 
 look at people based on their housing's relation to the environ- 
 ment. This is what we're trying to develop as a measure of 
 housing quality for 1980. 
 
 Barabba: Then we run into a problem. Let's say we come up 
 with a new measurement of housing quality and everybody agrees 
 that it's meaningful. We come out in 1980 and we say that 30 
 percent of the housing in the United States is not adequate. Is 
 that 30 percent an increase or a decrease from 1970? There's no 
 way that we can know because we are measuring it differently. 
 That's where the trade-off is: what is the value of this 
 information? Is it more important to get a new more meaningful 
 measurement than it is to be able to measure a possibly 
 meaningless change? 
 
 Mr. Levine: There is the same problem with rent. There are many 
 more supplemental rent programs now than there were 10 or 20 
 years ago. If we obtain a current measure of rent and compare it 
 with past measures, it may be meaningless, because we now have 
 subsidized rent in many areas. Therefore, you can't always look 
 at rent as a direct measure of quality in housing. In fact, you may 
 
 find, in some of the more difficult areas of the city, that the rents 
 are higher than they are elsewhere. Then rent becomes a different 
 measure. We have to evaluate our measurements in terms of what 
 I would call reality. 
 
 Student: Which still lays open the question about the head of a 
 household in a changing world. At what point does the term 
 become meaningless? For instance, I read recently that there are 
 something like 8 million unmarried female adults now in the 
 United States. Out of that 8 million let's say there are so many 
 million living alone that, perhaps, would check head of house- 
 hold. But what about those who are living with someone, either 
 male or female, in some type of egalitarian arrangement where 
 there are no children? 
 
 Mr. Levine: One way to cope with this problem is not to look at 
 a single measure. Let's take the case that you just mentioned of 8 
 million women living alone. One of the big explosions of the last 
 10 years has been the growth of the one-person household. We 
 look at size of household, size of family, and at the age of the 
 people who are living alone and find them at both ends of the 
 spectrum which tells us something. 
 
 It indicates that a lot more older people now are economi- 
 cally able to maintain their own households than 20 or 30 years 
 ago, before the advent of Social Security. 
 
 Secondly, it tells you that the social relationships which 
 existed 20 or 30 years ago of extended families within one 
 household no longer exist. The children don't live with their 
 grandparents, their parents, and all the way down the line. Our 
 society has changed so we look at size of household by age. We 
 look at the younger side and then at marital status and see that a 
 lot more women living alone are divorced. And we find by 
 looking at other measures that divorce is much easier to come by 
 and much more socially acceptable now than it has been. You 
 discover that today kids are economically able to leave their nests 
 much earlier. 
 
 The first thing to do is look at the whole spectrum of 
 available data. The social observer takes a look at all of the 
 component pieces. He looks at the income of these women, their 
 work experience, the number of children in the family. It's a 
 composite picture that you get by fitting individual pieces within 
 a social context. 
 
 The second thing is to recognize the problem, and 
 experiment during the decade as to how to change the questions 
 that will be asked in 1980. 
 
 Barabba: And if we find, as we're conducting these test 
 interviews, a lot of people who say, "I really resent the fact that 
 you have the head of the household question basically designed 
 with the male in mind," then that's something we've got to 
 consider, because if it's prevalent throughout a community it 
 could affect our response rate. 
 
 Student: You would be undertaking this not only as regards 
 identifying heads of household but other questions such as 
 income. 
 
 Mr. Levine: Sure, we have a lot of serious problems here. 
 
 Student: But that's important for economics, for welfare 
 purposes, Social Security purposes, Revenue Sharing. 
 
 114 
 
THE CENSUS FUNCTION 
 
 Mr. Levine: Very, very much so. 
 
 Mr. Barabba: So, in a sense, you have to develop a feedback 
 mechanism and that's what these tests are all about. 
 
 Student: Out of curiosity, I notice that you list 1860 as the 
 earliest possible year one could have been born. Is there anybody 
 in the United States over 1 10 years old? 
 
 Mr. Levine: We found too many centenarians, I'll tell you that. 
 Because we made an error in the way we set up the question on 
 the form, too many centenarians were reported. Seriously, 
 however, in answer to your question, I just don't know if 
 anybody in the United States is over 1 10 years old, which shows 
 you how confidential the data are. 
 
 Mr. Barabba: For statistical purposes, it really makes no 
 difference whether they're 110 or 120. It's immaterial. 
 
 Student: Could you explain how you caught the error on the 
 centenarians? 
 
 Mr. Levine: There are a lot of different ways we could have 
 caught it. The most direct would have been to have someone look 
 at the data and say, "My God, we goofed!" But that wasn't the 
 way we caught this one. 
 
 There is a technique called demographic analysis, which is 
 really how we measure the undercount in the census. Centenari- 
 ans are a relatively easy problem. What we can do is take the 
 number of people reported in 1960 who were 90 and over, then 
 age them. We have very good death records by age, so you apply 
 specific age mortality and that gives you a fairly good idea of how 
 many people there should be who are 100 years of age in 1970. 
 The one component you're missing is the migration of the 
 population. It's possible that a lot of people can leave the country 
 and go where they want to. But generally speaking, for that 
 population group 90 and over in 1960, there isn't a lot of 
 migration. 
 
 So we took a look at our demographic projection by 
 carrying the nonagenarian count forward, and decided there was 
 something amiss in our number of centenarians. Then we went 
 back and tried to find out why. When we looked at some 
 questionnaires and compared the write-in entry with the marked 
 entries it turned out that there had been some misunderstanding 
 on the part of the population. 
 
 Student: Do you do that for all the questions— try to catch 
 mistakes? 
 
 Mr. Levine: Basically, yes. We do it in terms of trying to measure 
 the undercount of the population. We do a tremendous amount 
 of work on the questionnaires in terms of trying to see how they 
 deviate from norms that, based on our experience, we would 
 expect them to look like. 
 
 For example, if you have a county in Kentucky, like Harlen 
 County, and you get the income distribution and you find out 
 the median income is $26,000, it doesn't take very long to decide 
 that there is a problem because Harlen County is basically a coal 
 mining community and the median income in 1969 was signifi- 
 cantly below the norm. We pull the Harlen County questionnaires 
 and may find that somebody spilled a cup of coffee on a 
 half-dozen questionnaires, and coffee will read on this question- 
 
 naire as a dot. Or we may find that the interviewer was very 
 poorly trained and made a mistake all the way along the line. We 
 found this to be the case in some of the rural counties where 
 farming is predominant, and we asked for net income and a 
 number of people reported gross income. 
 
 We compare our data against independent measures that 
 come from other sources, for example, against estimates of 
 various types of income developed from Social Security. We do 
 birth registration studies for kids under 5 to see whether we've 
 under-reported, overreported, or misreported them. You can do 
 internal analysis. If you see a household where you have a husband 
 26 and a wife 66 and three kids 10, 8, and 5, we have a feeling 
 that there may be something wrong there and you try to look at 
 it and make some judgments as to what problems we have. 
 
 Student: Is this all computerized? 
 
 Mr. Levine: Yes. But computerized or not, eventually you get 
 down to the human being who has to look at the data and make a 
 decision whether to change it or leave it alone. 
 
 Student: He's referring to tests. Do you have a program that 
 imputes this data? 
 
 Mr. Levine: Yes. We have edits; we impute data for people who 
 weren't counted; we have evaluation programs; we have post- 
 evaluation programs. After the census we spend a considerable 
 amount of money going back and doing reinterviews of a sample 
 to get some idea of what may have been wrong in the areas that 
 we measured originally. We ask more intensive types of questions 
 to find out what was wrong with various categories of questions. 
 
 Student: In other words, the 1970 census is not finished yet. 
 
 Mr. Levine: The 1970 census question is 99-percent finished, the 
 1 percent is what we're going to do in the way of further 
 evaluation relative to improving the 1980 census. Each census 
 contains within it the seeds of the next; things we've experi- 
 mented with or that we've learned. We'll go back and retabulate, 
 play around with it, look at the problem areas. 
 
 For example, one of the things we found in our demo- 
 graphic analysis was that, after several decades of improved 
 coverage of children under 10, this census seems to indicate 
 increased undercoverage of this group. It's very strange; we didn't 
 understand why. And we've been examining and exploring the 
 "why." You'll note that this questionnaire has eight lines listed 
 for people to report themselves. There's a suspicion that when the 
 household had more than eight people we may have lost them, as 
 the last people listed on most questionnaires are young children. 
 We have a feeling that some of our continuation sheets may have 
 been lost or separated from the original sheet and, thus, may have 
 been made separate households. The age of a household head 
 can't be under 10, so it would have been changed by imputation. 
 So we may have lost children that way. 
 
 There are two types of problems that you have in a census. 
 The first one is coverage of the population and the housing units. 
 Did you get to every person? Did you get to every unit? The 
 second question is, having arrived there, what types of content 
 problems did you have? Did they misunderstand? Did they report 
 the age incorrectly, did they misunderstand the race question? 
 You have content questions, identification-type questions. For 
 example, if you use a category such as "Mexican-Americans" and 
 
 115 
 
THE CENSUS FUNCTION 
 
 the population that is Mexican-American does not consider itself 
 as such but would only answer to "Mexicano," you have 
 introduced serious problems into your data. Therefore, we do a 
 content type of analysis and try to learn what we did wrong. 
 
 The question that I am asked most often is what do we do 
 the other 9 years? We don't just sit there and put sack cloth 
 and ashes on our heads and hide. We live in a very dynamic 
 society and it's becoming more dynamic and more complex every 
 year. We try to meet the data needs of the population at large, by 
 conducting sample surveys during the decade, on a wide variety 
 of topics. The Census Bureau conducts quite a number of surveys 
 on its own behalf, and those are surveys for which Congress has 
 mandated one way or another that we should receive the funding. 
 In other cases, the Census Bureau acts as a very large statistical 
 resource in which we undertake activities for other Federal 
 agencies that want statistical data compiled. What we try to do is 
 marry these two tasks in order to conduct them as efficiently as 
 possible. 
 
 Let me talk a little bit about what we do in between these 
 decennial censuses. As I've indicated, we don't really stop a 
 census anymore. In terms of its publication program we finished 
 the 1970 census in June 1973 and in July 1973 Congress gave us 
 money to start planning the 1980 census. However, let me go 
 back to what I consider the oldest, the most extensive, and 
 probably the most useful sample survey we have. The CPS, as we 
 affectionately call it, otherwise known as the Current Population 
 Survey, is a sample survey that came into being in August 1942. 
 It was developed basically to meet the question that I raised at 
 the beginning of my comments, namely, how many unemployed 
 are there in the United States? Obviously, unemployment was a 
 major question in the 1930's and sampling theory had reached a 
 point where it was feasible to develop a sample survey which 
 would meet three objectives: To produce reliable statistics each 
 month, to produce them quickly and efficiently, and for those 
 statistics to remain durable over time. Durable over time means 
 that as many times as you do it you get measures which are 
 consistent over time, that it doesn't blow up in your face. The 
 survey covers the total civilian noninstitutional population of the 
 United States, which is different from the total resident popula- 
 tion and from the total population. It's grown over time from an 
 original sample size of 25,000 households per month throughout 
 the United States, to where each month during the year we now 
 contact and interview some 45,000 households. 
 
 The CPS is probably the most accurate and reliable sample 
 in operation anywhere in the world. If you look at the 
 unemployment estimate it produces, it has a 3 percent coefficient 
 of error attached to it. This means that if the unemployment rate 
 is estimated at 5 percent, you have less than a two-tenths of a 
 percent variation. That's its accuracy. It's a very, very comprehen- 
 sive and extensive sample survey. 
 
 Barabba: We were talking with Dr. Wirthlin about that slope of 
 the confidence interval— how little each interview contributes 
 after you get to a certain number. In this case, if we use the CPS 
 as an example, we interview 45,000 people because two-tenths of 
 a percent variation is extremely important. 
 
 Mr. Levine: If all you are interested in knowing is how many 
 people use Brand X, you might not want to use— nor need to 
 use— 50,000 households. It's an expensive method. As the 
 Director has just said, many surveys for political and commercial 
 purposes deal with samples of 3,000 persons or less. In terms of 
 
 116 
 
 getting magnitudes of error where they're interested in 5 
 percentage points on a 50-percent base, they're willing to live 
 with a result which is accurate within a range of 45 to 55 percent. 
 In the CPS case a good deal of money, economic and social 
 policy, and human welfare rest on the data that come out of this 
 survey. That's why we've gone to the trouble of designing a 
 survey which has a high level of precision. 
 
 We collect the data every month by way of 1,000 to 1,200 
 interviewers spread throughout the country. The sample is 
 located in 461 places throughout the United States, including 
 Hawaii and Alaska— we have a sample in each State of the Union. 
 Each month during the week containing the 19th day of the 
 month, the interviewers knock on doors and begin contacting the 
 45,000 households to ask the questions that you have before you. 
 (See appendix.) The data relate to the week containing the 
 12th day of the month, so when an interviewer comes on Monday 
 of the week containing the 19th, he or she starts out by saying 
 "Last week what were you doing most of the week— Working, 
 keeping house, going to school, or something else?" 
 
 If you turn to the questionnaire and the yellow control 
 card (see appendix), I'll describe how we collect the 
 information. The control card contains a continuing record of the 
 basic classification information which the interviewer obtains at 
 the first visit to the household. We visit each household for 4 
 consecutive months, drop it out of the sample for 8 months, and 
 then visit it a year later for the same 4 months; thus, a household 
 participates in the Current Population Survey for a total of 8 
 months spread over a 16-month period. 
 
 This type of design has a number of advantages. First, if we 
 had to develop a new sample of 45,000 households each month, 
 the costs would increase appreciably. With this design, we bring in 
 25 percent of the households a month. Six thousand of those 
 households were last interviewed a year ago, 6,000 are coming in 
 for the first time, so each month some 75 percent of the 
 households are identical to those interviewed the month before 
 and 50 percent of the households are identical from year to year. 
 This approach is a compromise between obtaining the best 
 measure of change by dealing with the same sample and obtaining 
 the best measure of level by dealing with independent samples. 
 
 The data collected during the week of the 19th are 
 processed the following week, and 2 weeks from the time that 
 they're collected they're delivered to the Department of Labor 
 which is responsible for publishing these data on the following 
 Friday. So within 2-1/2 weeks from the time we collect 
 them, they're on the street. That's how urgent it is that we 
 distribute these data currently. 
 
 If you move over to page 4, column 2 of the CPS (see 
 appendix), you'll see that the second priority is people who 
 did not work last week but had a job or business from which they 
 were absent. Since they do have a job we find out whether they 
 were paid, the reason that they weren't working last week, 
 whether it is a full- or part-time job and their occupation and 
 industry. If you move over to the next column you'll see the 
 variety of information that we collect for the unemployed. What 
 they did to look for work, the number of weeks they were 
 looking for work, when they last worked, whether they're 
 looking for a full- or a part-time job and, finally, their last 
 occupation and industry so that we can classify the unemployed 
 by their skill level. 
 
 People answering "no" to those questions across the top 
 (20, 22, and 24) are neither working, absent from a job, nor 
 looking for work. These are the people who are not in the labor 
 
THE CENSUS FUNCTION 
 
 force, and we have characteristics in the extreme right-hand 
 column of the "not-in-the-labor-force" group by age, sex, marital 
 status, veteran status, and educational attainment. Those charac- 
 teristics apply to those in the labor force, as well. If you look 
 down the 24 series you'll also see that for those who are not in 
 the labor force, or in the so-called labor reserve, information is 
 obtained on when they last worked; if they worked in the last 
 5 years; their occupation and industry; whether they intend to 
 look for work; and why they don't look for work. We do collect a 
 very large variety of information on characteristics of the entire 
 population at work, looking for work, or not in the labor force. 
 
 Student: How are the data collected? 
 
 Mr. Levine: We ask the identical questions of every individual in 
 the household 14 years of age and over. It's true we generally get 
 the information from one adult for all members of the household. 
 That's how the data are collected. 
 
 The questionnaire you have is the March questionnaire and 
 what that tells you is that when we started this survey, the whole 
 focus was employment and unemployment, and we had a very 
 small questionnaire. It didn't take the planners, analysts, and 
 policymakers very long to figure out that if we were going out 
 and knocking on "n" thousands of doors every month to collect 
 the basic data, it would be very easy to ask another question, 
 another question, and another question. In addition to the basic 
 purpose, which still remains that of providing the labor force 
 data, we now ask supplemental questions virtually every month 
 of the year. 
 
 In March we conduct what we call our mini-census. In 
 effect, we get an intercensal measure every March, which is as 
 close as we can get to the April 1 date, of many of the same types 
 of data that we collected in the census. If you look at the 
 righthand pages 5, 7, and 9, you'll notice that in March we collect 
 information on work experience during the past year, migration, 
 family relationship, income, ethnicity, and we have an additional 
 series of questions which obtain the characteristics of the children 
 as well. Through the March supplement we can provide measures, 
 for families and individuals, of ethnicity, household composition, 
 income, educational attainment, a whole host of data. 
 
 Let me point out again that we are looking at this measure 
 from a variety of perspectives. One of the things we are looking 
 at, certainly, is the unemployment level. There's no doubt about 
 it. But the level alone is not too meaningful as you heard Mr. 
 Barabba mention this morning. To know that unemployment is 5 
 percent doesn't necessarily tell you anything, unless you have a 
 base of comparison. If you are conducting this survey, as we are 
 continually, and the movement into unemployment and out of 
 unemployment remains fairly constant, then your measure of 
 change is unaffected. 
 
 Really, what policymakers are looking at is not so much the 
 level as the change. When you pick up a headline and read that 
 unemployment is 5.2 percent, remember that isn't the only 
 statistic we produce. We produce monthly a tremendous amount 
 of detail on the labor market activity of the population, which 
 permits you to look beneath the overall number and examine 
 unemployment by marital status and race; how many are 
 teenagers; how many were laid off; how many have been looking 
 for work for short or long periods; what types of occupational 
 backgrounds they've had; which ones are in poverty; which ones 
 are not in poverty; which ones are primary earners in the family; 
 which ones are secondary earners. 
 
 What we do is try to refine and hone our concepts and 
 measurement techniques in order to minimize the variation 
 around the extremes. It's like any other measure. I will say that 
 over the years when you look at unemployment, and ex post 
 facto go back and look at turning points in economic activity— 
 when we're in a recession or when we're coming out of 
 one— you'll find that unemployment is a pretty good coincident 
 indicator. It moves very, very rapidly and very extensively. 
 
 Most policymakers are not just concerned with the overall 
 total. They look at what's happening to the heads of households, 
 the family composition, what's happening to women generally or 
 teenagers generally. Take the unemployment rate for teenagers, 
 for instance. Whereas the overall unemployment rate is now 
 about 5.2 percent, the teenager unemployment rate is running 
 close to 15 percent. Or, take the black rate which is twice the 
 white rate. These are the ratios to look at. 
 
 Mr. Barabba: Let's say that you look at this and somebody 
 wanted to change the unemployment question. How does a 
 person go about changing the questionnaire relative to unemploy- 
 ment? 
 
 Mr. Levine: I don't think "a person" could do it. These are 
 concepts which have been developed over long periods of time 
 after extensive research. Some of you may recall that when John 
 Kennedy became President in 1960 we were almost in a 
 recession. The Eisenhower Administration had come out of its 
 term of office with economic activity at a rather low level and 
 unemployment was rising rather rapidly— well over 6 percent. 
 When unemployment starts to rise, you've got a lot of people 
 raising questions about it. Reader's Digest, for example, pub- 
 lished a number of articles strongly suggesting that unemploy- 
 ment was really a meaningless concept. A new administration is 
 very sensitive to these kinds of changes and President Kennedy 
 appointed an advisory committee to look into the problem. 
 
 The Committee to Appraise the Employment and Unem- 
 ployment Statistics was under the chairmanship of Professor 
 Aaron Gordon of the University of California. The committee 
 examined the concepts and methods used in producing the labor 
 force estimates in great detail and issued its report and recom- 
 mendations to the President in 1962. Congress provided funding 
 in 1963. Extensive research and field testing was carried out 
 through 1965, before agreement was reached at all levels of 
 government on the changes to be made. Even then, both the old 
 and new concepts were carried out in separate surveys for all of 
 calendar 1966, both to provide a measure of the difference in 
 trend and level and the basis for restructuring retrospectively an 
 historical series by age, sex, marital status, and race. Finally, in 
 January of 1967 we shifted fully to the concepts, methods and 
 techniques that we use today. 
 
 Student: Did you find a significant difference between the 
 sample surveys that you were running during those years? I 
 assume that you did because you made a change. 
 
 Mr. Levine: There were a number of things that we changed 
 because the Gordon Committee asked us to change them. For 
 example, some of the fairly simple ones were: Prior to 1967 the 
 unemployment rate included everybody 14 years of age and over. 
 The Gordon Committee decided that 14 and 15 year olds, based 
 on the school enrollment rates in the United States, were 
 
 117 
 
THE CENSUS FUNCTION 
 
 inappropriate to include in this economic and social measure of 
 unemployment. They suggested that we collect the data, but that 
 it not be included in the official measure. The first thing was to 
 drop the 14 and 15 year olds from the official measure. 
 
 The second thing was that our unemployment questions 
 were not as specific as they are now. The unemployment concept 
 in the past was merely that you had looked for work within the 
 past 30 days but you didn't have to have an activity attached to it 
 nor did you have to be available for work. That immediately cut 
 out a lot of peripheral lookers. 
 
 Another thing we did on the employment side was to 
 tighten up some of our hour measures. We also shifted the 
 priorities around somewhat. Prior to the change, persons were 
 classified as looking for work even if they also had a job but had 
 not worked at the job during the survey week. The priorities were 
 changed at the suggestion of the Committee such that first 
 priority is given to those who are working, no matter whether 
 they were looking for work. Those with a job or business, 
 although not working during the survey week, are given second 
 priority; unemployment is the third. So there were some shifts of 
 that sort. 
 
 The result of some of these conceptual changes was to alter 
 the seasonal pattern, which is very closely observed. Before the 
 change you would find a big jump in unemployment in certain 
 months because kids who were about to get out of school were 
 looking for summer jobs. When we added the "availability for 
 work" concept, we moved those kids out of the unemployment 
 measure in May, for example, and moved them into the 
 unemployment measure in June. We shifted the temporal 
 relationships over the year because a lot of kids start to look 
 for jobs even though they're not available for work until early 
 June. Actually, we shifted the level in a very, very minor way, but 
 we changed the contexts, the movements within. 
 
 The difference between this survey and the census is that 
 the survey utilizes a skilled group of interviewers on whom we 
 spend a tremendous amount of time and effort training them in 
 how to handle relationship problems and general conceptual 
 problems. We can't have this detailed training, where we hire 
 185,000 interviewers to train and send out in a very short period 
 of time. 
 
 We have developed a highly skilled core of interviewers who 
 are very sensitive to the population's nuances. They are not only 
 sensitive to relationships, but race, ethnicity, marital status, 
 everything. We give them a rather intensive initial course of 
 training. We observe them frequently, and provide training every 
 month. We reinterview a sample of their work without their 
 knowledge as to when it's going to take place. We edit their work 
 every month and feed the results back to our Regional Offices to 
 determine whether their quality is acceptable or whether they 
 require additional training. Here we've put our resources into 
 people as well as into questionnaires. In the mail-back census we 
 have to look at ways of bridging the gap left because the 
 interviewer isn't there to help respondents interpret the forms. 
 
 Student: I have a question on "other income." I'm rather curious 
 as to why you don't list child support. I would assume that it's 
 shown under regular contributions when a person's not living at 
 home? 
 
 Mr. Levine: A moment ago, I mentioned that when we do this 
 survey we hold a group training session. That is, we bring our 
 interviewers together and we devote 8 hours to these concepts. 
 
 118 
 
 It's the home study, plus the group training, plus the 
 continual training that teaches the interviewer what's included 
 under income. The other point is that you can only get so much 
 on a questionnaire. You can't have 52 items on it. What we have 
 done is try to group different types of income in different 
 categories and regualr contributions do include child support. 
 
 Student: Except I think frequently you find that they're very 
 irregular. 
 
 Mr. Levine: You'll notice there's an item there for other income, 
 as well. We ask them for any income, for total income. It's a very 
 detailed probe in terms of income. 
 
 Student: Is it only every 10 years that you hire this 185,000? 
 How many of that 185,000 are more or less a year-round staff? 
 
 Mr. Levine: None. We have an independent field staff, the 
 numbers of which vary depending upon the workload. This 
 survey uses about 1,100. We probably have anywhere up to 1,700 
 interviewers working with us at any one point in time. And the 
 185,000 are hired only once every 10 years. 
 
 Student: I would assume a goodly portion probably did it 10 
 years ago and they're going to do it this time. 
 
 Mr. Levine: No, it's generally a completely new staff. 
 
 Student: Over what period of time are these 185,000 basically 
 employed? 
 
 Mr. Levine: This group will probably come on board sometime in 
 March of 1980 and they'll work through, possibly, June of 1980. 
 
 Let me mention some of the other types of information we 
 collect in the CPS. For the Labor Department, I've already 
 mentioned the data on annual work experience; we also report 
 multiple jobholdings, whether people are holding more than one 
 job; job mobility, how many different jobs people handle over a 
 year, whether they get premium pay and overtime pay; weekly 
 earnings; labor union membership; pension coverage; college 
 participation. For the Census Bureau itself, we put out reports on 
 migration, income, poverty, ethnicity, child spacing, fertility 
 expectations, education, school enrollment, family and household 
 characteristics, voting habits, housing characteristics, and vacancy 
 statistics. Then for what we call the nonrelatives, the people 
 outside the Labor Department and the Census Bureau, we've used 
 the CPS to collect information on health and disability, food 
 stamp participation, immunization status, television ownership, 
 farm wage workers, use of veterans' benefits, use of programs by 
 widows of veterans, recreation, participation in hunting and 
 fishing activities, to mention only a few. 
 
 The survey is very, very flexible. But one of the things that 
 happens is that the survey gets overburdened. We're visiting the 
 same people eight times and there's a limit to how many subjects 
 we can include. We have to come back, and we have to do it 
 quickly, trying to keep it to a 20-minute interview. Therefore, for 
 other Federal agencies, we do what we call ad hoc surveys, 
 totalling $50 million of work a year. 
 
 Some examples: The Census Bureau conducts the Health 
 Interview Survey for the National Center for Health Statistics. 
 This is a current weekly survey in which we measure the 
 morbidity status of the population, hospitalization, cost of health 
 
THE CENSUS FUNCTION 
 
 and medical care. For the Social Security Administration we've 
 done surveys of the aged, the disabled, and new beneficiaries of 
 different programs. We've done surveys on disabled veterans for 
 the Veterans Administration to enable them to decide whether 
 the Government benefits are adequate relative to today's living 
 costs. For the Bureau of Sport Fisheries and Wildlife of the 
 Interior Department, we've conducted surveys on recreation 
 fishing and hunting. 
 
 Student: Is the CPS a name survey or an address survey? 
 
 Mr. Levine: The CPS is a survey of households. We go to a 
 household for eight consecutive visits and if you live there for 
 3 months and then move, we pick up the family or individuals 
 that move in. We don't follow individuals in this particular 
 survey. 
 
 For OEO, when it existed, we did surveys on economic 
 opportunity, poverty, participation in the community action 
 programs, and opinions and attitudes towards government pro- 
 grams. As I mentioned, we conducted the Census Employment 
 Survey which was a post-censal survey in ghetto areas. 
 
 The surveys I've mentioned are all snapshots of points in 
 time. You can assimilate the analysis but you can't tell what 
 happened to Miss Jones. Therefore, for the Labor Department 
 and the Social Security Administration, we're doing what we call 
 longitudinal surveys. We've selected a sample of individuals. In 
 the case of the Manpower Administration, we're trying to look at 
 what happens to people at different cycles in the labor market. 
 For example, what happens to young people who are about to 
 move into the labor market for the first time? We're following a 
 sample of young women 30 to 44 years of age to see what 
 happens to them as they go through their fertility period, leave 
 the labor market, have their children, and come back into the 
 labor market. We're dealing with men in their prime age groups to 
 see what kind of mobility they have in the labor market. 
 
 We are doing surveys of people who are considerably older, 
 50 to 64, following them as they approach retirement, go through 
 the process of retirement and move into the full retirement phase 
 to see what kinds of social, psychological, and economic 
 problems they have. This is to enable the Social Security 
 Administration to develop legislation at the direction of the 
 Congress, and change benefit programs to meet the needs of the 
 population. 
 
 One major area that's relatively new for us is the area of 
 criminal victimization. As a result of a tremendous interest and 
 the intensity of concern which the Administration has had during 
 the past 5 years, and with the establishment of the Law 
 Enforcement Assistance Administration, the Census Bureau has 
 moved very rapidly to include measures on criminal victimization 
 in the household survey. Until the time this technique was 
 employed, most of the evaluation of criminal activity was based 
 on what's called the Uniform Crime Reports. These are basically 
 administrative statistics and are developed by getting measures 
 from each of the jurisdictions on the number of crimes reported. 
 
 There are a lot of problems in looking at administrative 
 statistics, particularly in the criminal area. We have the problem 
 that a lot of crimes are unreported to the police. Are they serious 
 or not? They can well be serious if you're trying to plan a 
 program developing measures to do something about the types of 
 crimes that aren't reported. Why aren't they reported? Are people 
 afraid of the police? That gives one insight immediately and 
 
 may necessitate change, perhaps, as to community relations, or 
 the perception of the police. 
 
 Secondly, are Uniform Crime Reports based on administra- 
 tive records at the mercy of the local police department? I'm not 
 trying to suggest that police departments are dishonest in any 
 way, shape, or form, but there are problems there. 
 
 A patrolman caught a kid running away after knocking over 
 a garbage can. One policeman may report it, the next policeman 
 may not. These are very serious matters of concern. 
 
 There are instances where you see the crime statistics 
 change because the police chief wants a larger budget and 
 so he may increase crime. He may tell his policemen to report 
 everything or he may change an administrative order. There are 
 different problems of reporting among different jurisdictions. 
 Some are more zealous in reporting exactly what's happening, 
 thinking they're a better police force. Another city may not have 
 the same view. And, of course, administrative reports do not 
 provide data on personal and family characteristics or other 
 victimizations that may have occurred. Although we ask about a 
 variety of events, only certain crimes were selected for measure- 
 ment. These are crimes against persons— rape, robbery and 
 assault; crimes against property-burglary, larceny, auto theft, 
 and robbery. 
 
 We are also trying to get a measure of business victimiza- 
 tion, commercial victimization. And we are conducting surveys of 
 inmates to find out what their socioeconomic characteristics were 
 before they became incarcerated, how many times they've been 
 incarcerated, what kind of work history they have. That, 
 incidentally, is one of the major concerns with using Uniform 
 Crime Reports as a source of data. We have a measure of criminal 
 activity but we know nothing about the social, psychological, and 
 economic impact of crime because the Uniform Crime Reports 
 don't give the age, sex, marital status, family status, residence, 
 length of time out of work, and the cost of victimization. 
 
 For the LEAA (Law Enforcement Assistance Administra- 
 tion), we're conducting continuing sample surveys of 70,000 
 households, over each 6-month period, to measure criminal 
 victimization. We're looking at inmate, prison and jail statistics. 
 We're doing surveys of the people in the criminal justice system 
 to get some centralized information on the experience, training, 
 and background of the people who are our law enforcement 
 officers, to find out what training they need and what kinds of 
 grants LEAA should give to local communities. We're doing 
 evaluation studies to measure the structure before and after 
 grants are given and programs implemented by local communities. 
 
 That gives some idea of the surveys we do. Let me mention 
 just briefly a few other items that the Census Bureau does, in the 
 other 9 years. We also produce population estimates and 
 projections. We provide estimates of the population each month, 
 its composition by age, sex, and geography. We provide popula- 
 tion estimates for States and counties, and now we're trying to 
 develop a methodology for providing current population esti- 
 mates for the 39,000 local governments which are benefited 
 under General Revenue Sharing, so that the apportionment 
 formula which is now based on 1970 data will be more current 
 and will reflect the fact that Los Angeles or Florida is growing 
 very rapidly, but some other State is not growing or is losing 
 population. 
 
 Student: Is there a bibliography or index stating what's avail- 
 able? 
 
 119 
 
THE CENSUS FUNCTION 
 
 Mr. Levine: The Census catalog. 1 There are a tremendous 
 number of reports, and the catalog gives you an indication of 
 what the population gets for the money they put into the 
 statistical system that we have developed. 
 
 I wanted to spend one moment to talk about a study that's 
 been developed. This is a new indicator. Indicators come and 
 indicators go. There are a lot of people around who say that 
 unemployment isn't the same type of indicator it was 20 years 
 ago. There are a lot more kids, marginal labor force members, 
 moving in. Twenty-five years ago you had the male member of 
 the household working and if he lost his job it was a very serious 
 situation. Now you have a lot of families where the husband and 
 wife both work. Kids are working. You have a lot of people who 
 are working full time, year round, but earning less than $3,000 
 which puts them in a poverty status. This is an attempt to 
 combine a series of data and attempt to take unemployment 
 solely out of the context of being the indicator that it was before 
 and tying it to a measure of need— trying to tie to unemployment 
 the number of people who are working, but working less than 
 full time for economic reasons, or who aren't earning a wage 
 above the poverty level, or who may have been discouraged from 
 looking. This is a new type of indicator which is now being 
 considered. 
 
 There are quite a lot of people who have played around 
 with what they call an Employment and Earning Inadequacy 
 Index. The EEI is the merger of a number of CPS data series to 
 produce a more meaningful measure of need with regard to the 
 social situation in which we find ourselves today. 
 
 The Bureau also is now very much involved in revenue 
 sharing, not just with General Revenue Sharing, but also with the 
 Comprehensive Employment and Training Act, and the Elemen- 
 tary and Secondary Education Act, because all of these formulas 
 inevitably contain somewhere in their legislation a little statement 
 that says the Census Bureau shall provide current measures of 
 whatever study is specified and we're frantically trying to gear up 
 to produce current measures with the reliability that's necessary 
 when money is to be distributed based on every single statistic. 
 
 It's a very difficult thing for the Bureau, and I think it's 
 going to cause the Bureau a lot of problems in the next 4 or 5 
 years as we try to meet the needs. Certainly our job would be 
 
 1 Bureau of the Census Catalog, available from the Superintendent of 
 Documents, U.S. Government Printing Office, Washington, D.C. 20402, or 
 any Department of Commerce district office. 
 
 a lot easier had Congress in its infinite wisdom decided to sponsor 
 a mid-decade sample survey or a mid-decade census. We are 
 working very diligently to try to convince the Congress to 
 authorize in 1985 a mid-decade census or large statistical survey. 
 Then, you'd have 5-year benchmarks, and the accuracy of 
 intercensal projections would be far improved than what we have 
 when we have to deal with a 10-year period during which 
 dramatic changes can occur, as the decade of the 1970's has 
 shown. 
 
 I'll close with two general comments. I touched only briefly 
 on the full scope of what the Census Bureau does. There are 
 many surveys, many activities, that we do on a day-to-day basis 
 that I haven't mentioned because of time constraints. But I do 
 want to make one final point. Having spent a good deal 
 of your time in telling you in effect that data are good 
 for you, I feel somewhat compelled to point out (and I hope you 
 don't consider me a traitor) that like medicine, more is not 
 necessarily good; and it could possibly be harmful. What I 
 mean by this is expressed in a statement I'd like to quote from a 
 book review that I read last week. This happened to be a book on 
 sex offenders, where the authors had compiled a tremendous 
 amount of statistical data. I thought the reviewer made a very 
 telling point, and one that should be kept in mind when using 
 data. He wrote, "The reader is assaulted by an awesome mass of 
 statistical detail but hardly any increase in understanding is likely 
 to emerge from this mountain of detail or this fallout of facts. 
 Facts can never speak for themselves, they must be made to 
 speak, and no amount of accumulation of data can substitute for 
 criteria which separate the trivial from the significant. There can 
 be no democracy of facts in a mature scientific enterprise." 
 
 What I'm saying is that everybody has facts and facts are 
 numbers on a piece of paper, but they don't mean anything until 
 somebody with intelligence analyzes them and puts them in a 
 meaningful context. They may not mean the same thing to 
 everybody, but they do tell a story, and somebody has to tell that 
 story. We can develop the mechanisms, we can produce the data, 
 but the user, the person who looks at those data, has to bring his 
 intelligence and analytical ability to the data as well. There are 
 different philosophies, different outlooks, and merely putting out 
 3 million tables doesn't solve the world's problems. 
 
 My point then, use statistics, look at statistics, but look 
 behind the data. First to see how they were produced, whether 
 they're reliable, whether they're based on good sound statistical 
 judgment, and then exercise your own native judgment. 
 
 120 
 
THE CENSUS FUNCTION 
 
 Page 1 
 
 UNITED STATES CENSUS 
 
 This is your Official Census Form 
 
 Please fill it out and mail it back 
 
 on Census Day, Wednesday, 
 
 April 1, 1970 
 
 Lai. 
 
 Ia2.!a3. ja4. 
 
 a5. 
 
 If the address shown above has the wrong 
 apartment identification, please write the 
 correct apartment number or location here: 
 
 How To Fill This Form 
 
 1. Use a black pencil to answer the questions. 
 
 This form is read by an electronic computer. 
 Black pencil is better to use than ballpoint or 
 other pens. 
 
 Fill circles "O" like this: • 
 
 The electronic computer reads every circle you 
 fill. If you fill the wrong circle, erase the mark 
 completely, then fill the right circle. 
 
 When you write an answer, print or write 
 clearly. 
 
 See the filled-in example on the yellow 
 
 instruction sheet. 
 
 This example shows how to fill circles and write 
 in answers. If you are not sure of an answer, 
 give the best answer you can. 
 
 If you have a problem, look in the instruction 
 
 sheet. 
 
 All instructions are numbered the same as the 
 questions on the Census form. 
 
 If you need more help, call the Census office. 
 
 You can get the number of the local office from 
 telephone "Information" or "Directory assistance. 
 
 U.S. Department of Commerce 
 Bureau ot the Census 
 Form D-60 
 
 3. Your answers are CONFIDENTIAL. The law 
 
 (Title 13, United States Code) requires that 
 
 you answer the questions to the best of your 
 
 knowledge. 
 
 Your answers will be used only for statistical 
 purposes and cannot, by law, be disclosed to any 
 person outside the Census Bureau for any 
 reason whatsoever. 
 
 The householder should make sure that the 
 information is shown for everyone here. 
 
 If a boarder or roomer or anyone else prefers 
 not to give the householder all his information 
 to enter on the form, the householder should 
 give at least his name, relationship, and sex 
 in questions 1 to 3, then mail back the form. 
 A Census Taker will call to get the rest of the 
 information directly from the person. 
 
 4. Check your answers. Then, mail back this 
 
 form on Wednesday, April 1, or as soon 
 
 afterward as you can. Use the enclosed 
 
 envelope; no stamp is needed. 
 
 Your cooperation in carefully filling out the form 
 and mailing it back will help make the census 
 successful. It will save the government the ex- 
 Dense of calling on you for the information. 
 
 / 
 
 PLEASE CONTINUE 
 
 121 
 
THE CENSUS FUNCTION 
 
 Page 1, con. 
 
 5. Answer the questions in this order: 
 
 Questions on page 2 about the people in your household. 
 Questions on page 3 about your house or apartment. 
 
 6. In Question 1 on page 2, please list each person who was living 
 here on Wednesday, April 1, 1970, or who was staying or visiting 
 here and had no other home. 
 
 EXPLANATORY NOTES 
 
 This leaflet shows the content of the 1970 census questionnaires. The content was 
 determined after review of the 1960 census experience, extensive consultation with 
 many government and private users of census data, and a series of experimental 
 censuses in which various alternatives were tested. 
 
 Three questionnaires are being used in the census and each household has an equal 
 chance of answering a particular form. 
 
 80 percent of the households answer a form containing only the questions on 
 pages 2 and 3 of this leaflet. 
 
 15 percent and 5 percent of the households answer forms which also contain the 
 specified questions on the remaining pages of this leaflet. The 15-percent form 
 does not show the 5-percent questions, and the 5-percent form does not show the 
 15-percent questions. On both forms, population questions 13 to 41 are repeated 
 for each person in the household but questions 24 to 41 do not apply to children 
 under 14 years of age. 
 
 The same sets of questions are used throughout the country, regardless of whether the 
 census in a particular area is conducted by mail or house-to-house canvass. An illus- 
 trative example is enclosed with each questionnaire to help the householder complete 
 the form. 
 
 122 
 
THE CENSUS FUNCTION 
 
 Page 2 
 
 DO 
 NOT 
 MARK 
 THIS 
 COL- 
 UMN 
 
 WHAT IS THE NAME OF EACH PERSON 
 
 who was living here on Wednesday, April 1, 1970 or 
 who was staying or visiting here and had no other home? 
 
 "\ / Head of the household 
 
 \ Print 1 Wife of bead 
 
 \ name! J Unmarried children, oldest first 
 
 \ ^ in this ) Married childen and their families 
 
 \ a; order I Other relatives of the head 
 
 \ !_j \ Persons not related to the head 
 
 2. HOW IS EACH PERSON RELATED TO THE 
 HEAD OF THIS HOUSEHOLD? 
 
 Pill one circle. 
 
 If "Other relative of head," also give exact relationship, for exam file, 
 mother-in-law, brother, niece, grandson, etc. 
 
 If "Other not related to head," also give exact relationship, for example, 
 partner, maid, etc. 
 
 \ 
 \ 
 \(T) Last i 
 
 \ 
 
 \ 
 \ 
 
 \ 
 
 \ 
 \ 
 
 1 
 
 First name 
 
 Middle initial 
 
 Head of household 
 
 O Wife of head 
 
 O Son or daughter of head 
 
 O Other relative i 
 
 of head— Print exact J 
 
 relationship -*- 
 
 Roomer, boarder, lodger 
 
 Patient or inmate 
 
 Other not related to head— Print exact 
 
 relationship 
 
 / 
 
 @ Last i 
 
 First name 
 
 Middle initial 
 
 O Head of household 
 O Wife of head 
 O Son or daughter of head 
 O Other relative 
 
 of head— Print exact \ 
 
 relationship -*- 
 
 Roomer, boarder, lodger 
 
 Patient or inmate 
 
 Other not related to head- Print exact 
 
 relationship 
 / 
 
 (S) Last i 
 
 First name 
 
 Middle initial 
 
 O Head of household 
 
 O Wife of head 
 
 O Son or daughter of head 
 
 O Other relative i 
 
 of head— Print exact | 
 
 relationship -»- 
 
 i i Roomer, boarder, lodger 
 i i Patient or inmate 
 ' ' Other not related to head— Print exact 
 relationship 
 
 / 
 
 \ 
 \ 
 \ 
 \ 
 \ 
 \ 
 "V 
 \ 
 
 \ 
 \ 
 \ 
 
 © 
 
 Last name 
 
 First name 
 
 Middle initial 
 
 O Head of household 
 O Wife of head 
 O Son or daughter of head 
 O Other relative p 
 
 of head— Print exact \ 
 relationship - 
 
 > Roomer, boarder, lodger 
 ' ' Patient or inmate 
 O Other not related to head— Print exact 
 
 relationship 
 
 / , 
 
 © 
 
 Last name 
 
 First name 
 
 Middle initial 
 
 Head of household 
 Wife of head 
 Son or daughter of head 
 Other relative 
 
 of head— Print exact \ 
 relationship ■ 
 
 Roomer, boarder, lodger 
 
 Patient or inmate 
 
 Other not related to head— Print exact 
 
 relationship 
 
 / , 
 
 
 h 
 
 
 @ Last i 
 
 First name 
 
 Middle initial 
 
 O Head of household 
 O Wife of head 
 O Son or daughter of head 
 O Other relative 
 
 of head— Print exact \ 
 relationship • 
 
 Roomer, boarder, lodger 
 
 Patient or inmate 
 
 Other not related to head— Print exact 
 
 relationship 
 
 / , 
 
 (?) Last i 
 
 First name 
 
 Middle initial 
 
 Head of household 
 Wife of head 
 Son or daughter of head 
 Other relative 
 
 of head— Print exact \ 
 
 relationship -*- 
 
 1 Roomer, boarder, lodger 
 O Patient or inmate 
 
 Other not related to head— Print exact 
 
 relationship 
 / , 
 
 1 . 
 \ 
 
 (T) Last i 
 
 First name 
 
 Middle initial 
 
 O Head of household 
 O Wife of head 
 O Son or daughter of head 
 C Other relative r 
 
 of head— Print exact j 
 relationship - 
 
 Roomer, boarder, lodger 
 
 Patient or inmate 
 
 Other not related to head- Print exact 
 
 relationship 
 
 / 
 
 9. If you used all 
 8 lines — Are there 
 any other persons 
 in this household? 
 
 Yes No 
 
 Do not list the others; we will 
 call to get the information. 
 
 10. Did you leave anyone out of Question 1 because 
 you were not sure if he should be listed — for 
 
 example, a new baby still in the hospital, or 
 a lodger who also has another home? 
 
 O Yes No 
 
 On back page, give name(s) 
 and reason left out. 
 
 123 
 
THE CENSUS FUNCTION 
 
 Page 2, con. 
 
 3. SEX 
 
 Fill one 
 circle 
 
 4. COLOR OR RACE 
 
 I I 
 
 Fill one circle. 
 
 If "Indian ( American)," also give tribe 
 If "Other," alio give race. 
 
 DATE OF BIRTH 
 
 5. Month and 
 year of birth 
 and age last 
 birthday 
 
 Print 
 
 Month 
 
 of 
 
 birth 
 
 Fill one 
 circle 
 
 7. Year of birth 
 
 Fill one circle 
 
 for first 
 three numbers 
 
 Fill one circle 
 for last 
 number 
 
 8. WHAT IS 
 EACH 
 PERSON'S 
 MARITAL 
 STATUS? 
 
 Fill one 
 circle 
 
 Male 
 
 Female 
 
 Male 
 
 Female 
 O 
 
 White 
 
 Negro 
 or Black 
 
 Japanese 
 
 Chinese 
 
 Filipino 
 
 Indian (Amer.) 
 ^ Print tribe -*- 
 
 Hawaiian 
 Korean 
 Other- Print 
 race 
 
 ' / 
 
 Month 
 Year 
 Age -_ 
 
 Jan. Mar. 
 Apr. June 
 July Sept. 
 Oct. Dec 
 
 186 
 187 
 188 
 189 
 190 
 191 
 
 White 
 
 Negro 
 or Black 
 
 Indian (Amer.) 
 Print tribe -»- 
 
 Japanese 
 
 Chinese 
 
 Filipino 
 
 Hawaiian 
 Korean 
 Other- Print 
 race 
 
 Month 
 Year _ 
 Age _ 
 
 Jan-Mar. 
 Apr-June 
 July-Sept. 
 Oct -Dec. 
 
 186 
 187 
 188 
 189 
 190 
 191 
 
 O 192 
 U 193 
 O 194 
 195 
 O 196 
 O 197 
 
 : 192 
 
 O 193 
 O 194 
 O 195 
 O 196 
 O 197 
 
 O 
 O 1 
 O 2 
 3 
 O 4 
 
 O 5 
 
 6 
 7 
 8 
 9 
 
 O 
 O 1 
 O 2 
 3 
 O 4 
 
 O 5 
 6 
 O 7 
 O 8 
 O 9 
 
 Now married 
 Widowed 
 Divorced 
 Separated 
 Never 
 married 
 
 Now married 
 Widowed 
 Divorced 
 Separated 
 Never 
 married 
 
 Make 
 
 no 
 mark 
 
 Ms 
 margin 
 
 Male 
 
 o 
 
 Female 
 
 Male 
 O 
 
 Female 
 
 White 
 
 Negro 
 or Black 
 
 Japanese 
 
 Chinese 
 
 Filipino 
 
 Indian (Amer.) 
 ^ Print tribe -*- 
 
 Hawaiian 
 Korean 
 Other- Print 
 race 
 
 t 
 
 Month . 
 Year _ 
 Age 
 
 Jan-Mar. 
 Apr. -June 
 July Sept 
 Oct -Dec. 
 
 186 
 187 
 188 
 189 
 190 
 191 
 
 O 192 
 193 
 O '194 
 O 195 
 
 196 
 O 197 
 
 O 
 1 
 
 O 2 
 3 
 
 O 4 
 
 O 5 
 
 6 
 Q 7 
 O 8 
 O 9 
 
 White 
 
 Negro 
 or Black 
 
 Indian (Amer.) 
 Print tribe -i 
 
 Japanese 
 
 Chinese 
 
 Filipino 
 
 Hawaiian 
 Korean 
 Other- Print 
 race 
 
 t 
 
 Month . 
 Year._ 
 
 Jan -Mar. 
 Apr. -June 
 July-Sept 
 Oct. Dec. 
 
 186 
 187 
 188 
 189 
 190 
 191 
 
 192 
 O 193 
 O 194 
 O 195 
 O 196 
 O 197 
 
 O 
 
 O 1 
 
 O 2 
 
 O 3 
 
 O 4 
 
 5 
 O 6 
 
 7 
 O 8 
 O 9 
 
 Now married 
 Widowed 
 Divorced 
 Separated 
 Never 
 married 
 
 Now married 
 Widowed 
 Divorced 
 Separated 
 Never 
 married 
 
 
 Male 
 
 Female 
 
 Male 
 
 Female 
 
 O 
 
 White 
 
 Negro 
 or Black 
 
 Indian (Amer.) 
 ^— Print tribe -) 
 
 Japanese 
 
 Chinese 
 
 Filipino 
 
 Hawaiian 
 Korean 
 Other- Print 
 race 
 
 Month 
 Year 
 
 Jan. -Mar 
 Apr. June 
 July-Sept 
 Oct -Dec. 
 
 186 
 
 187 
 188 
 189 
 190 
 191 
 
 O 192 
 O 193 
 O 194 
 195 
 O 196 
 O 197 
 
 
 O 1 
 O 2 
 O 3 
 O 4 
 
 O 5 
 6 
 O 7 
 O 8 
 O 9 
 
 White 
 
 Negro 
 or Black 
 
 Japanese 
 
 Chinese 
 
 Filipino 
 
 Hawaiian 
 Korean 
 Other- Print 
 race 
 
 Indian (Amer.) 
 Print tribe -+- 
 
 Month 
 Year . . 
 Age 
 
 Jan. Mar. 
 Apr. June 
 JulySept 
 Oct. -Dec. 
 
 186 
 187 
 188 
 189 
 190 
 191 
 
 O 192 
 O 193 
 O 194 
 O 195 
 O 196 
 O 197 
 
 O 
 
 O 1 
 
 O 2 
 
 O 3 
 
 O 4 
 
 O 5 
 
 O 6 
 
 O 7 
 
 8 
 
 O 9 
 
 Now married 
 Widowed 
 Divorced 
 Separated 
 Never 
 married 
 
 Now married 
 Widowed 
 Divorced 
 Separated 
 Never 
 married 
 
 Male 
 
 Female 
 O 
 
 Male 
 O 
 
 Female 
 
 White 
 
 Negro 
 or Black 
 
 Japanese 
 
 Chinese 
 
 Filipino 
 
 Hawaiian 
 Korean 
 Other- Print 
 
 race 
 
 Indian (Amer.) 
 _ Print tribe -»- 
 
 Month . 
 Year._ 
 
 Jan-Mar. 
 Apr. June 
 July-Sept 
 Oct. -Dec. 
 
 186 
 187 
 
 188 
 189 
 190 
 191 
 
 192 
 193 
 194 
 19b 
 196 
 197 
 
 
 
 
 
 o 
 
 5 
 
 o 
 
 1 
 
 o 
 
 6 
 
 
 
 2 
 
 o 
 
 7 
 
 o 
 
 3 
 
 o 
 
 8 
 
 Q 
 
 4 
 
 o 
 
 9 
 
 White 
 
 Negro 
 or Black 
 
 Japanese 
 
 Chinese 
 
 Filipino 
 
 Hawaiian 
 Korean 
 Other- Print 
 
 Indian (Amer.) 
 Print tribe -*- 
 
 Month 
 Year _ 
 Age 
 
 Jan-Mar 
 
 Apr. -June 
 
 O JulySept 
 
 Oct. Dec. 
 
 186 
 187 
 
 188 
 189 
 190 
 191 
 
 192 
 193 
 194 
 195 
 196 
 197 
 
 O 
 O 1 
 O 2 
 3 
 O 4 
 
 O 5 
 O 6 
 O 7 
 8 
 O 9 
 
 Now married 
 Widowed 
 Divorced 
 Separated 
 Never 
 married 
 
 Now married 
 Widowed 
 Divorced 
 Separated 
 Never 
 married 
 
 Mate 
 
 mark 
 
 this 
 margin 
 
 11. Did you list anyone in Question 1 
 who is away from home now — 
 
 for example, on a vacation or 
 in a hospital? 
 
 Yes 
 
 No 
 
 On back page, give name(s) 
 and reason person is away. 
 
 12. Did anyone stay here 
 on Tuesday, March 31, 
 who is not already 
 listed? 
 
 Yes 
 
 No 
 
 On back page, give name ol each 
 whom there is no one at his home 
 to report him to a census taker. 
 
 visitor lor 
 address 
 
 26:1 
 
 124 
 
THE CENSUS FUNCTION 
 
 Please answer question 
 10, 11, and 12 at the 
 bottom of page 2. 
 
 80, 15, ond 5 percent (100 percent) 
 
 Page 3 
 
 A. How many living quarters, occupied and vacant, are 
 at this address? 
 
 One 
 
 2 apartments or living quarters 
 
 3 apartments or living quarters 
 
 4 apartments or living quarters 
 
 5 apartments or living quarters 
 
 6 apartments or living quarters 
 
 7 apartments or living quarters 
 
 8 apartments or living quarters 
 
 9 apartments or living quarters 
 
 10 or more apartments or living quarters 
 This is a mobile home or trailer 
 
 Answer these questions for your living quarters 
 
 HI. Is there a telephone on which people in your living 
 quarters can be called? 
 Yes — >- What is 
 No the number? 
 
 Phone number 
 
 H2. Do you enter your living quarters — 
 
 O Directly from the outside or through 
 a common or public hall? 
 Through someone else's living quarters' 
 
 H3. Do you have complete kitchen facilities? 
 
 Complete kitchen facilities are a sink with piped 
 water, a range or cook stove, and a refrigerator. 
 • Yes, for this household only 
 
 Yes, but also used by another household 
 
 No complete kitchen facilities for this household 
 
 H4, How many rooms do you have in your living quarters? 
 
 Do not count bathrooms, porches, balconies, foyers, 
 halls, or half-rooms. 
 
 1 room 
 
 2 rooms 
 
 3 rooms 
 
 4 rooms 
 
 5 rooms 
 
 6 rooms 
 
 7 rooms 
 
 8 rooms 
 
 9 rooms or more 
 
 H5, Is there hot and cold piped water in this building? 
 
 Yes, hot and cold piped water in this building 
 No, only cold piped water in this building 
 O No piped water in this building 
 
 H6, Do you have a flush toilet? 
 
 Yes, for this household only 
 
 Yes. but also used by another household 
 
 No flush toilet 
 
 H7. Do you have a bathtub or shower? 
 
 Yes, for this household only 
 
 Yes, but also used by another household 
 
 No bathtub or shower 
 
 H8. Is there a basement in this building? 
 
 O Yes 
 
 O No, built on a concrete slab 
 
 No, built in another way (include mobile homes 
 and trailers) 
 
 H9. Are your living quarters — 
 
 Owned or being bought by you or by someone else 
 in this household? Do not include cooperatives and 
 condominiums here. 
 
 A cooperative or condominium which is owned or being 
 bought by you or by someone else in this household? 
 
 Rented for cash rent? 
 
 Occupied without payment of cash rent? 
 
 HlOa. Is this building a one family house ? 
 
 Yes. a one-family house 
 No. a building for 2 or more families 
 or a mobile home or trailer 
 
 b. // "Yes" — Is this house on a place of 10 acres or more, 
 - or is any part of this property used as a commercial 
 establishment or medical office? 
 
 Yes, 10 acres or more 
 
 Yes, commercial establishment or medical office 
 
 No, none of the above 
 
 Ml. 
 
 // you live in a one-family house which 
 
 you own or are buying — 
 
 What is the value of this property; that is, how much 
 
 do you think this property (house and lot) would sell for 
 
 if it were for sale? 
 
 Less than $5,000 
 $5,000 to $7,499 
 $7,500 to $9,999 
 
 $10,000 to $12,499 
 $12,500 to $14,999 
 $15,000 to $17,499 
 $17,500 to $19,999 
 
 $20,000 to $24,999 
 $25,000 to $34,999 
 $35,000 to $49,999 
 $50,000 or more 
 
 // this house 
 is on a place 
 of 10 acres or 
 more, or if 
 any part of 
 this property 
 is used as a 
 commercial 
 establishment 
 or medical 
 office, do 
 not answer 
 this question. 
 
 H12. Answer this question if you pay rent for your living quarters. 
 a. / / rent is paid by the month — 
 
 What is the monthly rent ? 
 
 Write amount here 
 
 and 
 
 Fill one circle 
 
 O Less than $30 
 i $30 to $39 
 $40 to $49 
 $50 to $59 
 
 O $60 to $69 
 O $70 to $79 
 
 $80 to $89 | 
 i $90 to $99 
 
 O $100to$119 
 ' i $120 to $149 
 O $150 to $199 
 
 $200 to $249 
 i $250 to $299 
 
 $300 or more 
 
 .00 (Nearest dollar) 
 
 If rent is not paid by the month — 
 
 What is the rent, and what period of time does it cover? 
 
 .00 per 
 
 (Nearest dollar ) 
 
 (Week, half -month, year, etc.) 
 
 FOR CENSUS 
 
 ENUMERATOR'S USE 
 
 ONLY 
 
 §4. Block 
 number 
 
 a 5. Serial 
 number 
 
 O O O 
 
 1 O O O 1 O O O 1 
 
 2 ' ■ ' ' 2 O O O 2 
 
 3 O O O 3 O O O 3 
 
 4 O O O 4 O O O 4 
 
 5 
 
 5 O O O 5 
 
 6 
 
 6 O O O 6 
 
 7 ' i 
 
 O 7 O O O 7 
 
 900 90009 
 
 B. Type of unit or quarters 
 
 Occupied 
 O First form 
 O Continuation 
 
 Vacant 
 Regular 
 
 Usual residence 
 _ elsewhere 
 
 Group quarters 
 O First form 
 Continuation 
 
 For a vacant unit, also fill 
 CD, A, H2 to H8, and 
 HlOloHU 
 
 C. Vacancy status 
 
 Year round — 
 
 For rent 
 
 1 For sale only 
 Rented or sold, not 
 
 occupied 
 O Held for occasional use 
 Other vacant 
 
 Seasonal 
 Migratory 
 
 Months vacant 
 ' Less than 1 month 
 • 1 up to 2 months 
 2 up to 6 months 
 6 up to 12 months 
 
 1 year up to 2 years 
 
 2 years or more 
 
 C/0 
 
 O 
 
 Make 
 no 
 mark 
 
 this 
 margin 
 
 125 
 
THE CENSUS FUNCTION 
 
 Page 4 
 
 15 and 5 
 percent 
 
 r 
 
 H13. Answer question H13 if you pay rent for your living quarters. 
 
 H19, Do you get water from — 
 
 
 In addition to the rent entered in H12, do you also pay for — 
 
 A public system (city water department, etc.) 
 or private company? 
 
 
 a. Electricity? 
 
 Yes, average monthly cost is -»- .00 
 
 An individual well? 
 
 
 Some Other source (a spring, creek, river, cistern, etc.): 
 
 
 No, included in rent Average monthly cost 
 
 ■ 
 
 
 O No, electricity not used 
 
 
 H20. Is this building connected to a public sewer? 
 
 b. Gas? 
 
 $ 
 j Yes, average monthly cost is -*- .00 
 
 O Yes, connected to public sewer 
 
 
 No, connected to septic tank or cesspool 
 
 
 No, included in rent Average monthly cost 
 
 No, use other means 
 
 
 No, gas not used 
 
 
 H21. How many bathrooms do you have? 
 
 c. Water? 
 
 
 $ 
 
 A complete bathroom is a room with flush toilet, bathtub or shower, 
 
 
 Yes, yearly cost is ^— .00 
 
 and wash basin with piped water. 
 
 
 j No, included in rent or no charge Yearly cost 
 
 A hidf bathroom has at least a flush toilet or bathtub or shower, 
 but does not have all the facilities for a complete bathroom. 
 
 
 d. Oil, coal, kerosene, wood, etc.? 
 
 $ 
 
 
 O No bathroom, or only a half bathroom 
 
 ' Yes, yearly cost is * .00 
 
 
 O No, included in rent Yearly cost 
 
 
 
 No, these fuels not used 
 
 O 1 complete bathroom 
 
 1 complete bathroom, plus half bath(s) 
 
 
 
 H14, How are your living quarters heated? 
 
 
 
 Fill one circle for the kind of heal you use most. 
 
 O 2 complete bathrooms 
 
 
 O Steam or hot water system 
 
 2 complete bathrooms, plus half bath(s) 
 
 
 ) Central warm air furnace with ducts to the individual 
 
 
 
 rooms, or central heat pump 
 
 O 3 or more complete bathrooms 
 
 
 O Built in electric units (permanently installed in wall, ceiling. 
 
 ■ 
 
 
 or baseboard j ^_ 
 
 
 H22. Do you have air-conditioning? 
 
 
 O Floor, wall, or pipeless furnace 
 
 Yes, 1 individual room unit 
 
 
 Room heaters with flue or vent, burning gas, oil, or kerosene 
 
 Yes, 2 or more individual room units 
 
 
 Room heaters without flue or vent, burning gas. oil. or 
 
 Yes, a central air-conditioning system 
 
 
 kerosene (not portable) 
 
 O No 
 
 
 O Fireplaces, stoves, or portable room heaters of any kind 
 
 
 
 
 
 H23. How many passenger automobiles are owned or regularly used 
 
 
 In some other way-Describe »- 
 
 by members of your household? 
 Count company cars kept at home. 
 
 
 
 None, unit has no heating equipment 
 
 None 
 
 1 automobile 
 
 H15, About when was this building originally built? Mark when the building 
 
 
 was first constructed, not when it was remodeled, added to, or converted. 
 
 O 2 automobiles 
 
 O 3 automobiles or more 
 
 
 O 1969 or 1970 O 1950 to 1959 
 
 
 
 O 1965 to 1968 O 1940 to 1949 
 
 ■ 
 
 ■ 
 
 
 1 1960 to 1964 '• 1939 or earlier 
 
 
 H16. Which best describes this building? 
 
 
 Include all apartments, flats, etc., even if vacant. 
 
 
 
 - A one-family house detached from any other house 
 
 
 
 A one-family house attached to one or more houses 
 
 
 
 A building for 2 families 
 
 
 
 A building for 3 or 4 families 
 
 
 
 A building for 5 to 9 families 
 
 
 
 ^ A building for 10 to 19 families ■ ™ 
 
 
 
 A building for 20 to 49 families 
 
 
 
 A building for 50 or more families 
 
 
 
 O A mobile home or trailer 
 
 
 
 Other— 
 
 
 
 Describe 
 
 
 H17. Is this building- 
 
 
 CD On a city or suburban lot? — Skip to H19 
 
 
 
 On a place of less than 10 acres? 
 
 
 
 - On a place of 10 acres or more? 
 
 
 H18. Last year, 1969. did sales of crops, livestock, and other farm products 
 
 
 from this place amount to — 
 
 
 
 Less than $50 (or None) | $2,500 to $4,999 
 
 
 
 $50 to $249 O $5,000 to $9,999 
 
 
 s. 
 
 $250 to $2,499 $10,000 or more 
 
 
 15 
 
 percent 
 
 126 
 
THE CENSUS FUNCTION 
 
 H24a. 
 
 r 
 
 How many stories (floors) are in this building? 
 
 1 to 3 stories 
 i 4 to 6 stories 
 — : 7 to 12 stories 
 
 
 \ 
 
 ' 13 stories or more 
 
 
 // 4 or more stories — 
 
 
 
 Is there a passenger elevator in this building? 
 
 Yes ■ No 
 
 
 H25a. 
 
 Which fuel is used most for cooking? 
 
 ■ 
 
 Coal or coke O 
 Wood O 
 
 
 1 From underground pipes 
 Gas serving the neighborhood. 
 ( Bottled, tank, or LP O 
 
 
 
 Other fuel . . O 
 
 b. 
 
 Fuel oil, kerosene, etc ~> 
 
 No fuel used O 
 
 Which fuel is used most for house heating? 
 
 ■ 
 
 Coal or coke C 
 
 
 1 From underground pipes 
 Gas serving the neighborhood . O 
 
 
 ' Bottled, tank, or LP O 
 
 Wood O 
 
 
 Electricity O 
 
 Other fuel . . O 
 
 c. 
 
 Fuel oil, kerosene, etc O 
 
 No fuel used O 
 
 Which fuel is used most for water heating? 
 
 ■ 
 
 Coal or coke O 
 
 
 I From underground pipes 
 Gas serving the neighborhood. 
 
 
 ' Bottled, tank, or LP O 
 
 Wood O 
 
 
 
 Other fuel . . O 
 
 
 Fuel oil, kerosene, etc O 
 
 No fuel used O 
 
 H26. 
 
 How many bedrooms do you have? 
 
 
 
 Count rooms used mainly for sleeping even if ust 
 
 d also for other purposes. 
 
 
 No bedroom O 3 bedrooms 
 
 
 1 bedroom 4 bedrooms 
 
 
 2 bedrooms ) 5 bedrooms or more 
 
 H27a. 
 
 Do you have a clothes washing machine? 
 
 Yes, automatic or semi-automatic 
 1 ' Yes, wringer or separate spinner 
 
 
 b. 
 
 • No 
 
 
 Do you have a clothes dryer? 
 
 
 
 ; ! Yes, electrically heated 
 
 
 c. 
 
 Yes, gas heated 
 
 No 
 
 
 Do you have a dishwasher (built-in or portable)'! 
 
 d. 
 
 • Yes No 
 
 
 Do you have a home food freezer which is separate from your refrigerator? 
 
 
 O Yes No 
 
 
 H2Sa. 
 
 Do you have a television set? Count only sets in working order. 
 
 
 Yes, one set 
 
 
 
 Yes, two or more sets 
 
 
 b. 
 
 No 
 
 
 // "Yes" — Is any set equipped to receive UHF broadcasts, 
 that is, channels 14 to S3? 
 
 
 Yes O No 
 
 
 H29. 
 
 Do you have a battery-operated radio? 
 
 Count car radios, transistors, and other battery-operated sets in working 
 
 
 order or needing only a new battery for operation 
 
 
 
 O Yes, one or more O No 
 
 
 H30. 
 
 Do you (or any member of your household) own a second home or other 
 living quarters which you occupy sometime during the year? 
 
 
 Yes O No 
 
 
 Page 5 
 
 5 
 percent 
 
 127 
 
THE CENSUS FUNCTION 
 
 Page 6 
 
 Tie 15-percent and 5-percent forms contain a pair of facing pages for each person m the househotd 
 (as listed on page 2). Shown on each pair of pages in the 15-perceat form art! fee questions des- 
 ignated as 15-percent here on pages 6, 7, ami 8. Shown oa each pair of pages m the 5-pe»ce*rt 
 form are the questions designated as 5-percent here or pages 6, 7, and 8. 
 
 15 and 5 
 percent 
 
 5 
 percent 
 
 15 
 percent 
 
 5 
 percent 
 
 15 ^ 
 
 percent 
 
 Name of person on line (T) of page 2 
 
 Last name 
 
 First name 
 
 Initial 
 
 13a. Where was this person bom? // born in hospital, give Stale o, 
 country where mother lived. If born outside U.S., see instruction 
 sheet; distinguish Northern Ireland from Ireland (Eire). 
 O This State 
 OR 
 
 (Name of State or foreign country; or Puerto Rico. Guam, etc.) 
 
 b. Is this person's origin or descent — (Fill one circle) 
 
 O Mexican O Central or South American 
 
 O Puerto Rican O Other Spanish 
 
 O Cuban O No, none of these 
 
 14. What country was his father born in? 
 
 O United States 
 OR 
 
 (Name of foreign country; or Puerto Rico. Guam, etc.) 
 
 15. What country was his mother born in? 
 
 United States 
 OR 
 
 (Name of foreign country; or Puerto Rico, Guam, etc.) 
 
 16. For persons born in a foreign country — 
 a. Is this person naturalized? 
 
 O Yes. naturalized _ 
 O No, alien ^ 
 
 Born abroad of American parents 
 
 20. Since February 1, 1970 , has this person attended regular 
 school or college at any time? Count nursery school, 
 kindergarten, and schooling which leads to an elementary 
 school certificate, high school diploma, or college degree. 
 
 O No fjg 
 
 O Yes, public 
 
 Yes, parochial 
 
 O Yes, other private 
 
 21. What is the highest grade (or year) of regular school 
 he has ever attended? 
 
 Fill one circle. If now attending, mark grade he is in. 
 Never attended school — Skip to 23 
 O Nursery school _ 
 O Kindergarten ™ 
 Elementary through high school (grade or year) 
 12 3 4 5 6 7 8 9 10 1112 
 
 000000 oo oooo 
 
 College (academic year) 
 
 1 2 3 4 5 6 or more 
 
 b. When did he come to the United States to stay? 
 
 1965 to 70 : 1950 to 54 \ 1925 to 34 
 
 O 1960 to 64 ! O 1945 to 49 1915 to 24 
 
 O 1955 to 59 ; O 1935 to 44 ' Before 1915 
 
 17. What language, other than English, was spoken in this 
 person's home when he was a child? Fill one circle. 
 O Spanish | O Other— 
 
 O French %"/? 
 
 O German O None, English only 
 
 18. When did this person move into this house (or apartment)? 
 
 Fill circle for date of last move. 
 
 1969 or 70 
 
 1968 
 
 1967 
 
 1965 or 66 
 1960 to 64 
 1950 to 59 
 
 1949 or earlier 
 Always lived in 
 this house or 
 apartment 
 
 19a. Did he live in this house on April 1, 1965? // in college or 
 
 Armed Forces in April 1%), report place of residence there. 
 
 : Born April 1965 or later | 
 
 r 
 
 Yes, this house 
 
 O No, different house 
 
 Skip to 20 
 
 b. Where did he live on April 1,1965? 
 
 (1) State, foreign country. 
 U.S. possession, etc. __ 
 
 22. Did he finish the highest grade (or year) he attended? 
 
 Now attending this grade (or year) 
 O Finished this grade (or year) 
 Did not finish this grade (or year) 
 
 23. When was this person bom? 
 
 O Born before April 1956 — Please go on with 
 
 questions 24 through 41. 
 
 O Born April 1956 or later- Please omit questions 24 through 
 41 and go to the next page 
 • | ; for the next person. 
 
 24. 
 
 // this person has ever been married — 
 
 Has this person been married more than once? 
 
 Once O More than once 
 1 -t- 
 
 When did he 
 
 get married? 
 
 When did he get married 
 for the first time? 
 
 (2) County _ 
 
 (3) Inside the limits of a city, town, village, etc.! 
 O Yes O No 
 
 (4) // "Yes," name of city, 
 
 town, village, etc. 
 
 Monlb Year Month Year 
 
 c. // married more than once — Did the first marriage end 
 because of the death of the husband (or wife)? 
 
 O Yes No — 
 
 25. // this is a girl or a woman — 
 
 How many babies has she ever 1234 5678 
 had, not counting stillbirths? OOOO OOOO 
 
 Do not count her stepchildren 
 
 or children she has adopted. 9 10 11 12 or None 
 
 more 
 
 OOOO o 
 
 26. // this is a man— 
 
 a. Has he ever served in the Army, Navy, or other 
 Armed Forces of the United States? 
 
 r— O Yes 
 | O No 
 
 b. Was it during — (Fill the circle for each period of service.) 
 
 Vietnam Conflict (Since Aug. 1964) O 
 
 Korean War (June 19i01ofan.l95S) O 
 
 World War II (Sept. 1940/ofuly 1947) O 
 
 World War I (April 1917 lo Nov. 1918) O 
 
 Any other time O 
 
 15 
 percent 
 
 IS and 5 
 percent 
 
 5 
 percent 
 
 15 and 5 
 percent 
 
 15 
 percent 
 
 J 
 
 128 
 
THE CENSUS FUNCTION 
 
 Page 7 
 
 5 
 
 percent 
 
 27a. Has this person ever completed a vocational training program? 
 
 For example, in high school; as apprentice; in school 
 of business, nursing, or trades; technical institute; 
 or Armed Forces schools. 
 v— O Yes O No — Skip to 28 
 
 b. What was his main field of vocational training? Fill one circle. 
 O Business, office work _ 
 
 Nursing, other health fields ^ 
 
 Trades and crafts (mechanic, electrician, beautician, etc.) 
 Engineering or science technician; draftsman 
 Agriculture or home economics 
 Other field- Specify ■ 
 
 ~} 
 
 28a. Does this person have a health or physical condition which 
 limits the kind or amount of work he can do at a job? 
 
 // 65 years old or over, skip to question 29- 
 O Yes 
 O No 
 
 b. Does his health or physical condition keep him from 
 holding any job at all? 
 
 O Yes 
 ■ O No 
 
 c. // "Yes" in a or b — How long has he been limited 
 
 in his ability to work? 
 
 O Less than 6 months 
 6 to 11 months 
 1 to 2 years 
 
 O 3 to 4 years 
 O 5 to 9 years 
 O 10 years or more 
 
 QUESTIONS 29 THROUGH 41 ARE FOR ALL PERSONS 
 
 BORN BEFORE APRIL ;9.56 INCLUDING HOUSEWIVES, 
 
 STUDENTS, OR DISABLED PERSONS AS WELL AS 
 
 PART-TIME OR FULL-TIME WORKERS 
 
 29a. Did this person work at any time last week ? 
 
 Yes- Fill this circle if this 
 person did full- or 
 part-time work. 
 (Count part-time work such 
 as a Saturday job. delivering 
 papers, or helping without 
 pay in a family business 
 or farm; and active duty 
 in the Armed Forces) 
 
 No- Fill this circle 
 if this person 
 did not work, 
 or did only 
 own housework, 
 school work, or 
 volunteer work. 
 
 Skip to 30 
 
 How many hours did he work last week (at all jobs)? 
 
 Subtract any time off and add overtime or extra hours worked. 
 O 1 to 14 hours O 40 hours 
 
 O 15 to 29 hours O 41 to 48 hours 
 
 O 30 to 34 hours O 49 to 59 hours 
 
 O 35 to 39 hours O 60 hours or more 
 
 Where did he work last week? 
 
 // he worked in more than one place, print 
 
 where he worked most last week. 
 If he travels about in his work or if the place does not 
 
 have a numbered address, see instruction sheet. 
 
 (1) Address (Number 
 and street name) 
 
 (2) Name of city, 
 town, village, etc.__ 
 
 (3) Inside the limits of this city, town, village, etc.? 
 
 O Yes 
 O No 
 
 (4) County 
 
 (5) State 
 
 (6) ZIP 
 Code 
 
 d. How did he get to work last week? Fill one circle for chief 
 
 means used on the last day he worked at the address given in 29c 
 
 O Driver, private auto 
 
 O Passenger, private auto 
 
 O Bus or streetcar 
 
 3 Subway or elevated 
 
 O Railroad 
 
 Taxicab 
 W.ilk. (I only 
 Worked at home 
 Other means— Specify 
 
 1 
 
 After completing question 20d, skip to question 33 
 
 30. Does this person have a job or business from which he was 
 temporarily absent or on layoff last week? 
 
 O Yes, on layoff 
 
 O Yes, on vacation, temporary illness, labor dispute, etc. 
 
 O No 
 
 31a. Has he been looking for work during the past 4 weeks? 
 
 Yes O No — Skip to 32 
 
 ri 
 
 Was there any reason why he could not take a job last week? 
 
 O Yes, already has a job 
 
 O Yes, because of tf|is person's temporary illness 
 
 O Yes, for other reasons (in school, etc ) 
 
 O No, could have taken a job 
 
 32. When did he last work at all, even for a few days? 
 
 O In 1970 | O 1964tol967 | O 1959or earlier 
 O In 1969 | O 1960tol963 \ O Neverworked 
 O In 1968 _ 
 
 Skip 
 to 36 
 
 15 
 percent 
 
 IS and 5 
 percent 
 
 — continued — 
 
 129 
 
THE CENSUS FUNCTION 
 
 Page 8 
 
 15 and 5 
 percent 
 
 5 
 percent 
 
 33-35. Current or most recent job activity 
 
 Describe clearly this person's chiej job activity or. business 
 last week, if any. If he had more than one job, describe 
 the one at which he worked the most hours. 
 
 If this person had no job or business last week, give 
 information for last job or business since 1960. 
 
 33. Industry 
 
 a. For whom did he work? // now on active duty in the Armed 
 Forces, print "AF" and skip to question 36. 
 
 { Name of company, business, organization, or other employer) 
 
 b. What kind of business or industry was this? 
 
 Describe activity at location where employed. 
 
 (For example: junior high school, retail supermarket, dairy farm, 
 TV and radio service, auto assembly plant, road construction) 
 
 c. Is this mainly — (Fill one circle) 
 
 O Manufacturing O Retail trade 
 
 O Wholesale trade O Other (agriculture, construction, 
 
 service, government, etc.) 
 
 34. Occupation 
 
 a. What kind of work was he doing? 
 
 ( For example: TV repairman, sewing machine operator, spray painter, 
 civil engineer, farm operator, farm hand, junior high English teacher) 
 
 b. What were his most important activities or duties? 
 
 (For example: Types, keeps account books, files, sells cars, 
 operates printing press, cleans buildings, finishes concrete) 
 
 c What was 
 his job 
 title? 
 
 35. Was this person— (Fill one circle) 
 
 Employee of private company, business, or 
 individual, for wages, salary, or commissions. . . 
 
 Federal government employee O 
 
 State government employee O 
 
 Local government employee (city, county, etc. ) . . . O 
 
 Self-employed in own business, 
 
 professional practice, or farm — 
 
 Own business not incorporated O 
 
 Own business incorporated O 
 
 Working without pay in family business or farm O 
 
 36. In April 1965, what State did this person live in? 
 
 O This State 
 OR 
 
 (Name of Stale or foreign country: or Puerto Rico, etc.) 
 
 GPO : in?2 O - 477-628 
 
 37. In April 1965, was this person — (Fill three circles) 
 
 a. Working at a job or business ( full or part-time)? 
 
 O Yes O No 
 
 b. In the Armed Forces? 
 
 O Yes C No 
 
 c. Attending college? ™ 
 
 O Yes O No 
 
 38. // "Yes" for "Working at a job or business" in question 37- 
 Describe this person's chief activity or business in April 1965. 
 
 a. What kind of business or industry was this? 
 
 b. What kind of work was he doing (occupation)? 
 
 c. Was he— 
 
 An employee of a private company or government agency. . . 
 Self-employed or an unpaid family worker O 
 
 39a. Last year (1969), did this person work at all, even for a few days? 
 
 Yes O No — Skip to 41 
 
 b. How many weeks did he work in 1969, either full-time or part-time i 
 
 Count paid vacation, paid sick leave, and military service. 
 
 O 13 weeks or less | | O 40 to 47 weeks 
 O 14 to 26 weeks i O 48 to 49 weeks 
 
 O 27 to 39 weeks O 50 to 52 weeks 
 
 40. Earnings in 1969 — Fill parts a, b, and c for everyone who 
 worked any time in 1969 even if he had no income. 
 
 (If exact amount is not known, give best estimate.) 
 
 a. How much did this person earn in 1969 1 
 
 in wages, salary, commissions, bonuses, /$ 
 
 or tips from all jobs? / — -- j-'- 
 
 ( (Dollars only) 
 
 (Before deductions for taxes, bonds, V 
 
 dues, or other items.) ) 0R None 
 
 b. How much did he earn in 1969 from his 
 own nonfarm business, professional 
 
 practice, or partnership? } 7--- 
 
 r • >~ r (Dollars only) 
 
 (Net after business expenses. If business ^^ \ 
 
 lost money, write "Loss" above amount.) I I ) 0R -' None 
 
 c How much did he earn in 1969 from 1 
 
 his own farm? / 
 
 (Net after operating expenses. Include earnings \ * .00 
 
 as a tenant farmer or sharecropper. If farm lost \ (Dollars only ) 
 
 money, write "Loss" above amount.) I Qp q None 
 
 41. Income other than earnings in 1969 — Fill parti a, b, and c. 
 
 (If exact amount is not known, give best estimate.) 
 
 a. How much did this person receive in I 
 
 1969 from Social Security or ( S® 
 
 Railroad Retirement? < DolUri °" l >> 
 
 1 OR O None 
 
 b. How much did he receive in 1969 from 
 public assistance or welfare payments? 
 
 Include aid for dependent children, old 
 
 age assistance, general assistance, aid 
 
 to the blind or totally disabled. \ — 'JTT, 7T — 
 
 \ ( Dollars only) 
 Exclude separate payments for mm 
 
 hospital or other medical care. J 0R None 
 
 c. How much did he receive in 1969 from \ 
 
 all other sources? I $ QO 
 
 Include interest, dividends, veterans' payments, ( ^Dollars only) 
 pensions, and other regular payments. I qd q |sj ne 
 (See instruction sheet.) ' 
 
 J 
 
 15 end 5 
 percent 
 
 5 
 percent 
 
 15 and 5 
 percent 
 
 130 
 
THE CENSUS FUNCTION 
 
 hi 
 
 , Q ■ u at trJT 
 
 < O 
 
 a. — '-' 
 
 
 j D 
 
 — o . 
 
 ry 
 
 r 3 l 
 
 r*i 
 
 M 
 
 01 5 2 o O 
 
 r=i 
 
 ■ ;-o J o^ > ; ; Q- 
 
 2 arm- 
 or a. «-; a: 
 OCL ^ ' 
 
 -JEtr 
 
 OOUJUJ 
 
 xxsm 
 
 : | z z 
 
 J£. m y < ^ ; 
 
 =) o. - 
 
 i I 
 
 S 5 0- u u o z 
 
 o r !± 
 
 S < Q 
 
 u + 
 
 s: uj x o - 
 
 t= E > * * 
 
 5 * o 
 
 : "5 : =E i ■* = - * - * r 
 
 'C/S3V /0 'Oft 3UU SIOl'J "V „ 
 
 laqujnu aun ~| | 
 
 s ^ s i o . 
 
 |S DDDDD 
 
 i § -. X. -I 5 Z 
 
 3 8 ™ E 
 
 r :-■=!? DDDD 
 
 
 
 < - > 
 a: jt : 
 
 < < : 
 
 if | E« 
 
 u q a 
 
 
 5 si * 
 
 
 
 
 
 
 
 - 2 £_2 
 
 Gl 1 i 1 
 
 V 
 
 J 
 
 a. < z 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 .' 
 
 
 
 ... 
 
 
 
 
 
 
 
 
 
 
 
 
 
 o 
 
 .--- 
 
 -) 
 
 n 
 
 —) 
 
 4: 
 
 ~> 
 
 6 
 
 X 
 
 X 
 
 X 
 
 X 
 
 X 
 
 :> 
 
 X 
 
 uDGuLG 
 
 131 
 
THE CENSUS FUNCTION 
 
 > 
 
 s 
 
 c 
 c 
 o 
 
 c 
 
 Q 
 Z 
 < 
 
 Z 
 
 O 
 Z 
 s> 
 
 3 
 
 r- 
 
 — UJ 
 
 ?2 
 
 c 
 
 ■r 
 
 r- 
 
 ti Control Card to 
 subsequent month s 
 4-month period of 
 ■asonolly vacant in 
 4-month periods of 
 
 
 
 
 Isil 
 
 : 
 r 
 O 
 
 E C 
 i J 
 
 n v 
 
 ; E 
 
 2 c 
 D !■ 
 
 3 
 -■ 
 
 5i 
 
 
 z 
 y 
 
 < 
 
 _l 
 
 U 
 
 t a separate unit, add 
 cuponts to thi s 
 ntrol Card: 
 
 eparate unit— Interview 
 n a separate Control 
 
 b- 
 
 O 
 
 X 
 
 
 
 X 
 
 O 
 
 I 
 
 
 -0 
 
 1 
 
 gj_ 6.2.= ! "•»'§ 
 
 <N 
 
 00 
 
 O u O ! */> O *— ' 
 
 ZOU j A 
 
 
 
 ft* ,- 
 E > 
 
 = X 
 
 C 
 
 "*" < o£ « ° „, £ » - 
 
 1/lllJ 5 Q -cS3=" t 'n = 
 
 < "i U n i-a Eg" 
 
 
 
 
 
 
 z i 1 
 
 
 z 
 
 z 
 
 Z 
 
 
 i<5 
 
 
 » ~ 
 
 00 
 
 
 3 
 
 
 s 
 
 
 
 z 
 
 
 
 z 
 
 
 
 Z 
 
 J ^ 
 
 Q. u 
 
 UJ ,_ S >-" — a.— 
 
 ■<■< aj— oj>o 00 
 U > _qc-c ;. ~ O- 
 
 . 3 
 
 5 1 "^ 
 
 ^ ;o 
 
 
 
 CC C 
 
 LU 
 
 iH*>. 
 
 s 
 
 
 
 
 
 o"-< 
 
 0^ 
 
 
 
 
 
 
 < a: 5 S - "> ° i 2 ,£ 
 
 or: u. E. 10 - |-c | ° 
 
 .t: a 
 
 : ° o 
 
 1- ^~ 
 E £ S ~" 
 
 E c -c = 
 
 
 
 1- -'" 
 a: 
 
 51 
 
 E - 
 
 
 V 
 
 >- 
 
 £ 
 
 
 
 
 
 a- o 
 
 2„a,ciol 6 
 
 ^ -c 
 
 
 U 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 E g- 
 
 Z 2 
 
 UL h-U 3 OJ «T 
 
 
 ^ s 
 
 
 t- 
 
 LU 
 
 a ^ 
 _i - 
 _j £ 
 < « 
 
 E c 
 * E 
 
 
 
 
 z 
 
 z 
 
 z 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ^|o OOO 
 
 
 
 
 
 u 
 
 
 ui E 
 
 
 
 
 
 
 
 
 
 
 
 30000 O 
 
 
 
 Tj c 
 
 
 < 
 
 a: 
 
 ■ 
 
 »" ° 
 
 r- 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 X 
 
 u 
 
 a: 
 
 
 
 
 
 
 
 
 
 
 
 c 
 o 
 
 
 
 
 
 ^ c 
 
 > 
 
 £ 
 
 Ijjl 
 
 
 0) 
 
 >- 
 
 >- 
 
 >- 
 
 
 
 
 
 o 
 
 
 
 
 
 
 LU 
 
 
 
 a 
 
 "° ' c 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 E = z i 
 
 
 
 
 
 Q 
 
 
 
 
 
 z: 
 
 z 
 
 z 
 
 
 
 
 
 E 
 
 Z 
 
 
 f>. 
 
 
 
 a 
 
 3 
 
 
 ul ^ "0 
 
 
 
 
 
 
 
 
 
 c j; 
 
 n 
 
 Eg.: 5 
 
 
 z ■? 
 
 
 
 < 
 
 
 Q 
 
 c ^"~ a. 
 
 
 
 
 
 
 
 
 
 * o 
 
 S2 
 
 s-5^£ ■» 
 
 
 j !; 
 
 
 
 Q 
 
 
 zd 
 
 Q.-S" i 
 
 
 Cf) 2 
 
 o 2 
 
 o 2 
 
 
 
 
 
 
 X 
 
 z 
 
 x^.E 
 
 
 ~*£ 
 
 
 c 
 
 LU 
 
 
 Q- 
 
 Z; 
 U 
 
 
 ■O 
 
 " 
 
 Oi 
 
 
 C3i 
 
 
 
 
 
 Z E 
 
 
 
 
 
 
 
 
 u 
 
 
 
 
 
 
 
 
 
 
 4, = 
 Q. c 
 
 £ 
 O 
 
 O 
 
 
 
 _o * 
 
 ■ - in O -J 
 
 h- 
 < 
 
 
 
 
 0. -£5— o. 
 
 
 1 "D 
 
 '■a 
 
 ■O 
 
 
 
 
 
 J £ O OOO 
 
 c*° 
 
 cc 
 
 
 
 
 C 3 c 
 
 
 
 
 
 
 
 
 
 
 
 
 
 °T3 
 
 E o 
 
 0> o_ 
 
 O 
 
 • -^I ^ O OOO 
 
 ^ g m >o r^ 4» 
 
 
 11 
 
 D O 
 
 ■£ o ~o = 
 
 Z 
 
 2 
 
 1- 
 
 
 
 O 
 
 z 
 
 z 
 
 
 
 z 
 
 
 
 
 
 E_c 
 
 CO 
 
 Z> 
 
 ■/> *" ^ 
 E c - E 
 
 O D O 
 
 2 -" 
 
 
 _c "° 
 
 
 < 
 
 z 
 2 
 
 O^o 5>o. ^ 
 
 10 
 
 
 
 
 
 
 
 
 = E 
 
 
 X^ ^ CO 
 
 
 r^ 
 
 a> c 
 
 UJ 
 
 -u-°2 qJ^Q. 
 
 
 V! 
 
 U] 
 
 Ul 
 
 
 
 
 
 D __ 
 
 -o-g 
 
 = s 
 
 
 E =| lOOOO 
 
 
 
 
 l- 
 
 ily-: ui 
 
 
 > 
 
 > 
 
 >- 
 
 
 
 
 
 
 
 UJ 
 
 Q 
 
 
 
 >. 
 
 
 
 S w 
 
 
 1^^ 2 OOO O 
 
 
 
 1 c 
 
 £ 
 
 5" 
 
 *• 
 
 
 ~o 
 
 -0 
 
 ~D 
 
 
 
 
 
 5» "^ o 
 
 
 ci j _ 
 
 
 
 UJ 
 
 
 c 
 
 C 
 
 c 
 
 
 
 
 
 > - 1 
 
 x _ X 
 - E-g 
 
 
 
 
 
 
 1- 
 
 a; _o 01 
 
 
 
 
 O 
 
 
 
 
 
 
 
 
 
 
 
 .r 00 
 
 3 
 
 E x -0 
 
 CO - ° CN 
 
 
 c 
 
 _o 
 
 c 
 
 
 
 
 
 
 >- j •> 
 uj ^ ^ 
 > f! 
 
 a- c o- 
 
 3 ° 0, 
 «/> E > 
 
 >■ =-8 
 
 s » 
 
 7, ^ 
 
 0000 
 00 00 
 
 1— 
 Z 
 
 E 
 p- 
 
 c 
 
 
 
 O 
 
 O 
 
 Z 
 > 
 
 <o i; ° E 
 
 "> -O Q_ 
 
 5 
 
 E 
 01 
 
 -0 
 
 E 
 ~o 
 
 E 
 ~o 
 
 
 
 
 
 z §- o 
 
 
 
 
 
 
 O 
 u_ 
 
 UJ 
 
 I! 
 
 E 
 
 1 
 
 | 5 
 
 
 O 
 
 O 
 
 O 
 
 
 
 
 
 < in 
 
 > ■/> - 
 
 D 3 3 
 
 >» u 
 
 
 
 
 '. <u 
 
 ^1 
 
 1 -*"! 
 
 LU 
 
 -J 
 CQ 
 
 ;_ ^ ^ J 
 
 • 
 
 
 D 
 
 □ 
 
 D 
 
 
 
 
 
 o 5 2 
 
 3 ° 
 E Z c 
 
 J "> 
 
 
 
 r 
 
 • E 
 
 
 — 
 
 3 D O *" 
 
 «i c — 
 
 < 
 
 g — T -O 
 
 
 | 
 
 J 
 
 
 
 
 
 
 S IN HOU 
 
 scribe to H 
 ds of enume 
 
 ^_|o _ ^ ^ J, 2 
 
 
 < 
 
 "£ D 
 
 
 
 iJ-loi 1 
 
 b 
 
 _l 
 
 _l 
 
 _i 
 
 
 
 
 
 — 
 
 O 
 Z 
 
 -O 
 
 UJ « 3 
 
 > >^i z 
 
 
 • • 
 
 
 J 
 
 J 
 
 J 
 
 :i 
 
 
 
 
 OOOOOOO 
 
 •"" c o 
 
 5 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 z 2 ;: 
 
 
 3 
 
 
 
 
 
 
 
 
 
 
 ,, 
 
 
 
 
 3 l-S. 
 
 
 
 
 
 
 
 Q- 
 
 J 
 
 
 
 
 
 
 
 
 
 n 
 
 
 
 
 Q_ j. JZ 
 
 
 
 
 
 
 
 
 
 t 
 
 
 
 
 
 
 
 
 
 ;'; 
 
 
 
 
 O c "£ 
 
 wi 
 
 
 
 
 
 
 
 
 3 
 
 
 
 
 
 
 
 
 
 
 
 
 
 u_ p, o 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 a E 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 > 4 
 
 O 
 
 
 
 
 
 
 
 
 z 
 
 lO O O 
 
 
 
 
 
 
 
 
 ,,.' 
 
 
 
 
 c 
 
 i_ *— j- T O O O — 
 
 
 
 
 CN 
 
 
 
 
 
 
 
 
 , 
 
 
 
 
 
 
 i * 
 
 o^ - 1 m -*t n ,_ 
 * -° oOO-CKC^O- 
 
 
 5 
 
 ;o 
 
 
 
 f.. 
 
 
 
 
 
 ' 
 
 
 
 
 
 >»4, — 1 — >— -— ■ — 1— V 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 tl 
 
 o= 5£££°°5 
 
 
 3 
 
 
 
 
 S 2 
 
 
 
 
 
 
 
 
 
 
 
 "> " ounooooo 
 
 
 w 
 
 
 
 
 
 
 •- 
 
 — ^ "~ 
 
 
 
 
 
 
 
 
 
 E ^_ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 c «- c^c^c^o-c^c^C^ 
 
 
 3 
 
 
 
 
 
 
 3 
 
 ; g 8 
 
 
 
 
 
 J. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 LL 
 
 
 
 
 
 
 
 
 
 ■* CJ 
 
 J3 
 
 
 ,5 
 
 
 
 
 
 
 O 
 
 ° .5-^ 
 
 _ 
 
 
 
 
 
 
 
 
 *> ■£ 
 
 
 
 5 
 
 
 
 
 
 
 O 
 
 °" ^ rvj 
 
 ' J 
 
 
 
 
 i 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 * 
 
 ~° C 
 
 a .- 
 "-J c 
 
 loo 
 3oo 
 
 E . . 
 
 
 a 
 
 
 
 
 
 < 
 
 u 
 
 
 9) ^ ° 
 
 
 
 
 
 j 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 *-° ~ 
 
 
 
 -a 
 
 
 
 
 
 ^ c o> 
 
 
 
 
 
 
 
 
 
 T ^ 
 
 -* c 2 
 
 
 ~o 
 
 
 ~a *- 
 
 
 
 * Uj aJ 
 
 
 
 
 
 ;' 
 
 
 
 X ^ 
 
 O = VJ g 
 
 
 "D 
 
 
 "s * 
 
 
 
 
 
 
 
 
 g 
 
 
 
 
 E "" 
 
 
 
 c 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 fz 
 
 U E- E 
 
 
 
 
 >• T 
 
 
 
 
 
 
 
 
 
 
 
 
 * c 
 
 _ ^1 — 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 => 
 -O x: 
 
 > ,^ a. ** *- 
 
 
 O 
 
 U 
 
 4> O* O 
 
 >- z z 
 
 
 
 
 
 
 
 
 I 
 
 ^ 
 
 
 
 c 
 
 c ^, c • 
 
 
 - 
 
 
 
 
 
 
 
 
 
 ■:r- 
 
 l 
 
 
 £ 
 
 — _ -J — 
 
 
 — 
 
 
 
 
 
 
 
 
 h 
 
 O 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 O 
 
 O 
 
 
 
 
 "" 
 
 
 r " 
 
 
 
 J z 
 
 ' 
 
 
 CN 
 
 m 
 
 z 
 
 u. 
 
 
 
 132 
 
THE CENSUS FUNCTION 
 
 INTERVIEWER CHECK ITEM 
 
 FORM CPS-1 
 
 U.S. DEPARTMENT OF COMMERCE 
 
 CONTROL NUMBER 
 
 
 Only CPS-I for household Q 
 
 
 Bureau of the Census 
 
 
 
 First CPS-I of continuotion h'hold... Q 
 
 
 
 
 
 Second CPS-1 of continuotion h'hold . [ 
 
 
 
 
 
 Third, fourth, ond 5th CPS-1 Q 
 
 
 
 
 
 
 Form Approved 
 
 O.H.B. No. 41-R1202-14 
 
 PSU SEGMENT SERIAL 
 
 LINE NO.OFH'HOLD RESP. 
 
 
 
 
 
 NON H'HOLD RESPONDENT Q 
 
 
 (Specify o/vj Send Intercom*) 
 
 
 
 
 
 \ 
 
 
 
 
 
 INTERVIEW 
 
 
 
 
 
 ANY ENTRY IN CPS "I . yl — 1 
 
 ITEMS 23A- E ' 5LJ 
 
 ,NoQ 
 NONINTERVIEW 
 
 
 
 
 
 
 (f* n fi fp) fp) b [Ni ff 
 
 
 
 TYPE A Q 
 
 
 ^ W In LH £ LRI LI 
 
 
 
 TYPE B □ 
 
 
 
 
 
 TYPEC ■ Q 
 
 
 
 
 
 8JWHY 
 
 I^SCtMl DDM 
 
 133 
 
THE CENSUS FUNCTION 
 
 25. LINE NO. 
 
 26. RELATIONSHIP TO HEAD OF HOUSEHOLD 
 
 (Control Card Item 14b) 
 
 (Er.tr, rtlotior.ir.tp 
 circle below) 
 
 Child 
 
 Other relative . . . 
 
 Nonrelotive ■■ 
 own rels. in 
 household 
 
 Nonrelotive - no 
 own relatives 
 m household . 
 
 OFFICE USE ONLY 
 
 Family No. 
 I £ 3 "f 5 G 
 
 Fam. Rel. 
 
 Child 
 
 Other 
 relative 
 
 Type 
 
 Sec. I 
 Sec. F 
 Sub. F 
 
 27. AGEfHarto 
 circle only) 
 
 9 O 
 
 10 O 
 
 11 
 12 
 
 13 O 
 
 White 
 Negro 
 Other 
 
 Male.. 
 
 Female 
 
 e e 
 
 : i 
 
 3 £ 
 3 3 
 
 26. RELATIONSHIP TO HEAD OF HOUSEHOLD 
 
 (Control Cord Item lib) 
 
 (Enter relations 
 and nark one 
 circle below} 
 
 Child 
 
 Other relative . . 
 
 Nonrelotive ■■ 
 own rels. in 
 household . . . 
 
 Nonrelotive ■■ no 
 own relatives 
 in household. 
 
 OFFICE USE ONLY 
 
 Family No. 
 
 1831-5 
 
 Fom. Rel. 
 Child 
 
 Type 
 
 Sec. I 
 Sec. F 
 Sub. F 
 
 27. AGEi-Korio 
 circle only) 
 
 9 O 
 
 10 
 
 n : 
 
 12 o 
 
 13 : 
 
 29. RACE 
 
 Wh.te 
 Negro 
 Other : 
 
 Male.. 
 Female 
 
 a 
 
 -j 
 
 X 
 
 u 
 
 o 
 z 
 
 s 
 
 aj 
 
 25. LINE NO. 
 
 O 
 
 1 I 
 
 3 3 
 
 
 6. RELATIONSHIP TO HEAD OF HOUSEHOLD 
 
 (Control Cord Item lib) 
 
 
 27. AGE fHV* on. 
 
 circle only) 
 
 3 7 
 
 1 O 8 '1' 
 
 2 O 9 
 
 3 O 10 
 
 4 O 11 
 
 5 12 
 
 6 13 O 
 
 ■ 
 
 29. RACE 
 
 White 
 Negro 
 Other 
 
 
 (Enter relationship 
 and mark one 
 circle below) 
 
 
 
 m 
 
 Child 3 
 
 OFFICE USE ONLY 
 
 
 ■ 
 
 Family No. 
 
 30. SEX 
 
 Male. . 
 Female 
 
 Other relative ... 
 
 Nonrelotive - 
 own rels. in 
 household .... 
 
 Nonrelotive -- no 
 own relatives 
 in household . . 
 
 Fom. Rel. Type 
 
 ch,id : i Sec. i : 
 
 Other $ec. F : 
 relotive C Sub. F : 
 
 1 
 
 25. LINE NO. 
 
 
 
 1 I 
 
 26. RELATIONSHIP TO HEAD OF HOUSEHOLD 
 
 (Control Card Item lib) 
 
 (Enter relations 
 and mark one 
 circle below) 
 
 Child 
 
 Other relative . . . 
 
 Nonrelotive •• 
 household 
 
 Nonrelotive •- no 
 own relatives 
 in household . , 
 
 OFFICE USE ONLY 
 
 Fomily No. 
 
 Fom. Rel. 
 
 Child 
 
 Other 
 
 Type 
 
 Sec. I 
 Sec. F 
 Sub. F 
 
 27. AGEi*"lo 
 
 c.rcl, only] 
 
 a o 
 
 1 
 
 2 O 
 
 3 O 
 
 4 O 
 
 5 O 
 6 
 
 9 O 
 
 10 O 
 
 11 O 
 
 12 O 
 
 13 O 
 
 29. RACE 
 
 White 
 Negro 
 Other 
 
 Male. 
 Female 
 
 O 
 
 : i 
 
 3 3 
 
 26. RELATIONSHIP TO HEAD OF HOUSEHOLD 
 
 (Control Cord Item 14b) 
 
 (Ente, 
 
 Child 
 
 Other relative . . . 
 
 Nonrelotive -- 
 own rels. in 
 
 household 
 
 Nonrelotive -- no 
 own relatives 
 m household.. 
 
 OFFICE USE ONLY 
 
 Family No. 
 
 I 
 
 G 
 
 Fam. Rel. 
 
 Child 
 
 Other 
 
 relotive 
 
 Type 
 
 Sec. I 
 Sec. F 
 Sub. F 
 
 O 
 
 27. AGEiHorlbo 
 circle only) 
 
 7 
 8 
 9 
 
 10 O 
 
 11 O 
 
 12 O 
 
 13 O 
 
 White 
 Negro 
 Other 
 
 30. SEX 
 
 25. LINE NO. 
 
 
 
 1 I 
 E £ 
 3 3 
 
 26. RELATIONSHIP TO HEAD OF HOUSEHOLD 
 
 (Control Card Item 14b) 
 
 
 cle below) 
 
 Child 
 
 Other relotive . . 
 
 Nonrelotive - 
 own rels. in 
 household . . . 
 
 Nonrelotive - no 
 own relatives 
 in household . 
 
 OFFICE USEONLi 
 
 Fa 
 
 I £ 
 
 ily No. 
 
 Fom. Rel. 
 
 Child 
 
 Other 
 relative 
 
 Type 
 Sec. I 
 Sec. F 
 Sub. F 
 
 27. AGE (Mark a 
 circle only) 
 
 9 O 
 
 10 o 
 
 11 
 
 12 
 13 O 
 
 White O 
 Negro O 
 Other O 
 
 30. SEX 
 
 Male.. 
 Femole 
 
 25. LINE NO. 
 
 i : 
 
 26. RELATIONSHIP 
 
 (Control Cord Item 
 
 TO HEAD OF HOUSEHOLD 
 
 14b) 
 
 (Enter 
 
 ship 
 
 cle below) 
 
 Child 
 
 Other relotive . . . 
 
 Nonrelotive -- 
 own rels. in 
 
 household 
 
 Nonrelotive •• no 
 own relatives 
 in household . . 
 
 OFFICE USE ONLY 
 
 Family No. 
 I £ 3 "r 5 6 
 
 Fam. Rel. 
 
 Child 
 
 Other 
 relative 
 
 Type 
 Sec. I 
 Sec. F 
 
 Sub. F 
 
 27. AGE (Hark o 
 Circle only) 
 
 9 O 
 
 10 O 
 
 11 O 
 
 12 O 
 
 13 O 
 
 29. RACE 
 
 White 
 Negro 
 Other O 
 
 Male.. 
 Femole 
 
 25. LINE NO. 
 
 C 
 
 1 I 
 £ £ 
 3 3 
 
 26. RELATIONSHIP TO HEAD OF HOUSEHOLD 
 
 (Control Card Item 14b) 
 
 Child 
 
 Other relative . . . 
 
 Nonrelotive ■- 
 
 household 
 
 Nonrelotive -- no 
 own relatives 
 in household . . 
 
 OFFICE USE ONLY 
 
 Fomily No. 
 I 5 3 "r 5 6 
 
 Fam. Rel. 
 
 Child 
 
 Other 
 relative 
 
 Type 
 Sec. I 
 Sec. F 
 Sub. F 
 
 27. ICE (Hark o 
 orcle only) 
 
 2 
 3 O 
 
 29. RACE 
 
 White 
 Negro 
 Other 
 
 Male.. 
 Female 
 
 134 
 
THE CENSUS FUNCTION 
 
 I. INTERVIEWER CHECK ITEM 
 
 Only CPS-1 lor household O I (Fill all applicable 
 
 First CPS-1 of continuation h'hld.. O \ Hems on this page) 
 
 Second CPS-I of continuation h'hld. O / (Transcribe Items 2-13 
 
 Third, fourth, etc. CPS-1 O \ from first CPS-1) 
 
 FORM CPS-1 
 12 20 73 
 
 26.1 1 
 
 COMMERCE 
 
 £ TO 8 UNIT [?©l?arjlLMJi£)K) SiWiY 
 
 MARCH 1974 
 
 O.U.B. No. 41-R120214 
 
 2. SAMPLE |3. CONTROL NUMBER 
 
 A C 
 
 MONTH 
 
 YEAR 
 
 O 
 
 0. INTERVIEWER CODE 
 
 ABCDEFGH JKLM 
 
 o o o o o o o o o o o o 
 - -J 
 
 O I c 3 •- 3 S 2 8 ' : 
 
 11. DATE COMPLETED 
 
 2. LINE NO H HLD RESP 
 
 Non. h'hld.resp. (Specify) O (Send Inter Comm) 
 
 4. TYPE OF LIVING QUARTERS 
 HOUSING UNIT 
 
 House, apartment, dot 
 
 HU in nontronsient hotel, motel, etc 
 
 HU, permonent.in transient hotel, motel, etc. 
 
 HU in rooming house 
 
 Mobile home or trailer 
 
 HU not specified obove (Describe below). . . . 
 
 Uescnbe belt 
 } 
 
 OTHER UNIT 
 
 Quarters not 
 
 HU in rooming or boarding house 
 
 Unit not permanent 
 
 in transient hotel, motel, etc. . . 
 
 Tent site or trailer site 
 
 5. LAND 
 USAGE 
 
 (Transcribe 
 
 from Control 
 Cord uems 
 10a, II or 12) 
 
 A O 
 
 B O 
 
 C O 
 
 D O 
 
 E O 
 
 Other not HU (Describe below) . 
 t 
 
 6. PSU NO 
 
 
 
 I I 1 
 
 6 G 
 
 ; ? 
 
 7. SEGMENT 
 NO. 
 
 I I I 1 
 
 3 3 3 
 
 O- q_ o- 
 
 SERIAL 
 NO. 
 
 
 
 1 : 
 E c 
 3 }: 
 
 G G 
 
 ? i 
 
 9. HOUSE- 
 HOLD 
 NO. 
 
 13. TYPE INTERVIEW 
 
 Non interview C 
 
 Personol .... O 
 
 Tel. -regular O 
 
 Tel.-collback O 
 
 ICR filled... O 
 
 NONINTERVIEW 
 
 14. (Mark reason and 
 
 race.) 
 
 REASON 
 
 RACE OF 
 HEAD 
 
 No one home 
 
 | 
 
 Temporarily 
 dbsent . O 
 
 White C 
 
 Refused ... O 
 
 Negro C 
 
 Other-Occ. O 
 (Describe 
 below) • 
 
 Other C 
 
 TYPE B 
 
 Vacant - reyulor 
 
 Vocont - storoge ot h'hld furniture 
 Temp. occ. by persons with URE - 
 
 Unfit or to be demolished 
 
 Under construction, not ready .... 
 Converted 
 
 to temp, business or storage ■ • 
 Occ. by Armed Force 
 
 members or persons under 14 . . 
 
 Unoccupied tent site or trailer site 
 Permit granted, 
 
 construction not started 
 
 54. INTERVIEWER CHECK ITEM 
 
 a. Is any person classified as one of the 7 Spanish 
 origins in Item 52 ? 
 
 ( Mexican-American, Chicano, Mexican (Mexicano), 
 Puerto Rican, Cuban, Central or So. Amer. or 
 Oth»r Spamsh) 
 
 Other (Specify below) . 
 
 (Fill 
 
 16) 
 
 (Omit 
 
 16-17) 
 
 TYPE C 
 
 Demolished 
 
 House or trailer moved . 
 
 Outside segment 
 
 Converted to permonent 
 business or storoge. 
 
 Merged 
 
 Condemned 
 
 Built after April 1,1970 
 
 Other (Describe below) . 
 
 r 
 
 (Omit 
 16-17) 
 
 SEASONAL STATUS 
 
 Migratory workers. 
 Seosonal 
 
 (FillHVSif 
 
 HV in Item4) 
 
 {(Fill Item 17 belo 
 [if HO m Item *) 
 
 17. Is this unit usually occupied 
 Summers only .... O I '' 
 
 Winters only 
 Other 
 
 as instructed 
 on back of 
 Control Card.) 
 
 (Describe below) 
 \ 
 
 Ves 
 
 (Ask b) 
 / 
 
 'I: 
 
 (Co 
 
 b. Did the head of this household .... live at this 
 address during the week of November 19, 1973? 
 
 Yw C No 
 
 ( Co to page 2) 
 
 55. CPS-581 required for one or more persons on this document . 
 List line numbers and item numbers below: 
 
 Line Numbers) 
 
 Item Number(s) 
 
 CODER NUMBER 
 
 ABCDEFGH JKLM 
 
 oooooooooooo 
 
 0183 "r 56? 8 9 
 
 135 
 
THE CENSUS FUNCTION 
 
 18. LINE NUMBER 
 
 19. Whot was . . . doing most of 
 
 LAST WEEK- 
 
 ^ Working 
 
 Keeping house 
 ' Going fo school 
 or something else? ^ 
 
 Working (Skip to 20A) WK C 
 
 With a job but not at work J C 
 
 Looking for work LK C 
 
 Keeping house H 
 
 Going to school S 
 
 Unable to work (Skip to 24) U - 
 
 Retired R C 
 
 Other (Specify) OT O 
 
 20C. Does . . . USUALLY work 35 
 
 hours or more o week at this |ob? 
 
 Yes I What is the reason . . . 
 worked less than 35 
 hours LAST WEEK' 
 
 No What is the reason . . . 
 
 USUALLY works less 
 than 35 hours a week? 
 (Mark the appropriate reason) 
 
 Slack work 
 
 Material shortage 
 
 Plont or machine repair G 
 
 New job storted during week . C 
 
 Job termmoted during week . . O 
 
 Could find only part-time work O 
 Holiday (Legal or religious) . . 
 
 Labor dispute O 
 
 Bod weother O 
 
 Own illness O 
 
 On vacation O 
 
 Too busy with housework, 
 
 school, personal bus., etc. O 
 
 Did not wont full-time work . . C 
 Full-time work week 
 
 under 35 hours O 
 
 Other reoson (Specify) O 
 
 \ 
 
 (Skip to 23 and enter /ob worked 
 at last week) 
 
 20. Did . . . do ony work ol oil 
 LAST WEEK, not counting 
 work oround the house' 
 (Note . // farm or business 
 operator in hh.. ask about 
 unpaid work) J 
 
 Yes No O (Gou2lj 
 
 20A. How mony hou 
 did . . . work 
 LAST WEEK 
 otoll |obs? 
 
 20B. INTERVIEWER 
 CHECK ITEM 
 
 (Skip 
 
 49+ 
 1-34 
 
 i 23,1 
 (Co to 20C) 
 
 35-48 
 
 ,l.„ ,,. Mill 
 
 20D. Did . . . lose ony time or 
 toke ony time off LAST 
 WEEK for ony reoson 
 such os illness, holiday 
 or slock work? 
 
 Yes 
 
 How mony hours 
 did . . . 
 take off 
 
 (Cor reel 20A if lost lime 
 not already deducted, 
 i(20A reduced below 3S, 
 correct 20B and (ill 20Ct 
 otherwise, skip to 23.) 
 
 Did . . . work ony 
 overtime or at more than 
 one job LAST WEEK' 
 
 How mon 
 hours did , 
 
 work' 
 
 .Ira 
 
 (Correct 20A and 208 , 
 necessary if extra hours 
 not already included and 
 skip to 23.) 
 
 (Skip to 23) 
 
 21. Ill I in 19, skiptollA.) 
 Did . . . have a lob or 
 business from which he 
 wos temporanlv absent or 
 on lovoff LAST WEEK' 
 
 Yes 
 
 (Co lo 22; 
 
 21A. Why wos . . . absent 
 
 from work LAST WEEK' 
 
 Own illness.... O 
 
 On vocation .... O ^* 
 
 Bod weother . . 
 
 Labor dispute . . 
 
 New |ob to begin 
 within 30 days 
 
 Temporary layoff 
 (Under 30 days} 
 
 Indefinite loyoff 
 (30 days or more or 
 no del. recall dale) 
 
 Other (Specify). . 
 
 o 
 
 (Skip to 
 
 I) 22B and 
 
 22C2) 
 
 IfSkip lo 
 \ 23C3) 
 
 21B. Is . . . getting wages or 
 salory for any ol the 
 time off LAST WEEK' 
 
 Yes O 
 
 No 
 
 Self-employed 
 
 21C. Does . . . usually work 
 35 hours or more o week 
 at this job? 
 
 Yes C 
 
 (Skip to 23 and e 
 held last week) 
 
 
 OFFICE USE ONLY 
 INDUSTRY OCCUPATION 
 
 '. 
 P 
 
 Q O 
 R O 
 
 s o 
 
 T 
 
 u 
 . 
 
 A 
 
 X o 
 
 Y 
 
 z 
 
 
 A 
 
 i : 
 
 B 
 
 G c 
 
 C O 
 
 3 3 
 
 . 
 
 o- o r 
 
 E 
 
 
 F O 
 
 6 G 
 
 G O 
 
 
 H C 
 
 
 J 
 
 
 K O 
 
 
 L .. 
 
 '/ 
 
 M O 
 
 G G & 
 
 ? ? 
 
 Ref. 
 
 22. (If LK in 19. skip to 22A.) 
 
 Has . . . been looking for wo 
 during the past 4 weeks' 
 
 Yes _■ No 
 
 22A. What has . . . been doing in the 
 last 4 weeks to find work? (Mark 
 all methods used; do not read list.) 
 
 Checked 
 with— pub.employ.ogency ' 
 
 pvt. employ. ogency ' 
 
 employer directly 
 
 friends or relatives 
 
 Placed or answered ads 
 
 Nothing (Skip to 24} I 
 
 Other (Specify in notes, e.g., MDTA, , 
 union or prof, register, etc.) 
 
 22B. Why did . . . start looking for ■ 
 
 work' Was it because . . . lost ^ 
 or quit o job at that time (pause) 
 or was there some other reason? 
 
 Lost job O 
 
 Quit job O 
 
 Left school O 
 
 Wonted temporary work O 
 
 Other (Specify in notes) Q 
 
 22C. 1) How many weeks 
 hos . . . been 
 looking for work' 
 
 2) How many weeks 
 ago did . . . start 
 looking for work' 
 
 3) How many weeks 
 ago was . . . laid 
 
 o 
 
 1 I 
 
 3' D 
 
 G G 
 
 : : 
 
 22D. Has . . . been looking for full-time 
 or port-time work? . 
 
 22E. Is there any reason why . . . could 
 not toke a ,ob LAST WEEK? 
 
 Yes i Already hos a job O 
 
 \ Memporary illness .... O 
 
 JGoing to school O 
 
 No O (Other (Specify i 
 
 s) 
 
 22F. When did . . . last work at a 
 
 full-time |ob or business losting 
 2 consecutive weeks or more? 
 
 1969 Or later (Write month and v 
 
 tr), . 
 
 (Month and year) 
 
 Before 1969 j 
 
 Nev. worked full-ti me 2 wks. or more O 
 Never worked at oil O 
 
 (Skip to 23 and enter last full-time 
 civilian fob lasting 2 weeks or more, job 
 from which laid off, or "Never Worked") 
 
 24. INTERVIEWER CHECK ITEM | 
 
 .... I 
 
 Unit in rotation group: 
 
 (Mark one circle only) 
 
 . 2, 3, 4, 6, 7 or 8 (End questions) 
 
 I, or 5 (Co to24A) 
 
 24A. When did . . . last work for pay at a 
 regulor |ob or business, either full- or 
 port-time? 
 Within past 12 months 
 
 1 up to 2 yeors ogo. . 
 
 2 up to 3 yeors ago. . 
 
 3 up to 4 years ago. . 
 
 4 up to 5 years ago. . 
 
 5 or more years ogo . 
 Never worked ( 
 
 1 (Co to 24B) 
 
 .st. F 
 
 24B. Why did . . . leave that job' 
 
 Personal, family 
 
 find, pregnancy) or school O 
 
 Heolth O 
 
 Retirement or old age C 
 
 Seasonal job completed Q 
 
 Slock work or business conditions O 
 Temporary 
 
 nonseasonal |ob completed ... O 
 Unsatisfactory work 
 
 arrangements (Hours, pay, etc.) O 
 
 Other 
 
 24C. Does . . . want a regular job now, 
 either full- or part-time? 
 
 Maybe-if depends (Co 
 
 (Specify in notes) \ 
 
 No 
 
 Don't kn 
 
 [(Skip to24E) 
 
 24D. What are the reasons ... is not 
 looking for work? 
 
 (Mark each reason mentioned) 
 • Believes no work ovoiloble 
 
 in line of work or area 
 
 Couldn't find ony work Q 
 
 Locks nee, schoolin g, 
 
 training, skills or experience C 
 f n i '. yers 
 
 think too young or too old ... O 
 Other 
 
 pers. handicap in finding job O 
 
 Can't arrange child care O 
 
 Family responsibilities O 
 
 In school or other training O 
 
 III heolth , physical disability . . . O 
 
 Other (Specify in notes) C 
 
 Oon't know C 
 
 24E. Does . . . intend to look for work 
 of any kind in the next 12 months' 
 
 Yes 
 
 It depends (Specify i 
 
 No 
 
 Oon't know 
 
 (If entry in 24B, describe ;ob in 23) 
 
 23. DESCRIPTION OF JOB OR BUSINESS 
 
 23A. For whom did . 
 
 *Ork' (Name of company, bu. 
 
 mployer) 
 
 23B. Whot kind of business or industry is this' (For example: TV and radio mfg.. 
 
 ■ Labor Depl., farm.)* 
 
 23C. Whot kind of > 
 
 ] 23E. Wos this person 
 
 An employee of PRIVATE Co., 
 
 bus., or individual for woges, salary or comm. 
 
 A FEDERAL government employee 
 
 A STATE government employee 
 
 A LOCAL government employee 
 
 . doing? (For example, 
 
 col engir 
 
 stock clerk, typist, far 
 
 Self-empl. in OWN bus., prof, practice, or forr 
 , \ Yes... 
 
 Is the business incorporated? 
 
 .F O 
 
 . S 
 
 -L 
 
 I 
 
 23D. What were . . .'s most important octivities or duties? 1F01 
 operates printing press, finishes concrete.) 
 
 mple: types, keeps < 
 
 I No lor farm) .... SE 
 
 Working WITHOUT PAY in fam. bus. or form . 
 NEVER WORKED 
 
 .. WP 
 .NEV 
 
 25. LINE 
 NUMBER 
 
 25a. POPULA 
 TION STATUS 
 
 Forces 
 member 
 
 IAF m Control 
 Card item 23) 
 
 26. RELATIONSHIP TO HEAD OF 
 HOUSEHOLD (Enter relationship 
 
 in box and mark one circle below) 
 
 Head with other relotives 
 (mcl.wi/e) inh'hld O 
 
 Head with no other 
 relotives in h'hld O 
 
 Wife of head 
 
 Child 
 
 Other relative of heod 
 
 Non-rel. of heod with own 
 
 rels. fuel, wife) in h'hld. , 
 Nonrelotive of head with no 
 
 own relatives in h'h Id. . . 
 
 OFFICE USE 
 ONLY 
 
 Fam. No. 
 
 ' Fam. Rel. 
 Head . . . C 
 Wife.-.. C 
 Child... 
 Other rel. C 
 
 Type 
 Sec. I . . . 
 Sec. Fam. 
 Sub. Fam. 
 Pri. I . . . 
 
 29. RACE 
 
 White 
 
 Negro 
 
 Other 
 
 28. MARITAL STATUS 
 
 Married — civilian 
 spouse present 
 
 Morried - AF 
 spouse present 
 
 Married - 
 
 spouse absent 
 in Armed Forces 
 
 Morried — 
 
 spouse absent 
 other reasons . 
 
 Widowed 
 
 Divorced 
 
 Separated 
 
 Never morried . . . 
 
 30. SEX AND 
 
 VETERAN STATUS 
 Male 
 
 Vietnom Ero . O 
 Korean War . . O 
 World War 1 1 
 World War 1. . 
 Other service O 
 Nonveferon . . O 
 
 Femole O 
 
 31. HIGHEST GRADE 
 ATTENDED 
 
 E H C 
 [ I 1 
 c c c 
 3 3 3 
 
 32. GRADE 
 COM- 
 PLETED 
 
 136 
 
THE CENSUS FUNCTION 
 
 33. INTERVIEWER 
 CHECK ITEM 
 
 Civil.cn 14* 
 
 (Ask 34) 
 
 Armed Forces 
 member 
 
 34. In 1973 how mony weeks did ... work 
 either full iime or port time not 
 counting work around the house' 
 (Include ,..,., vacation and paid sick leave 
 
 eks) 
 
 If M9 fSi.pIo 38) 
 If 50-52 (Skip la 40) 
 
 35. Even though ... did not work in 1973, did 
 he spend any time trying to find a |ob 
 
 (Ask 
 
 Yes items No 
 
 36-37) 
 
 (Skip I 
 
 36. How mony different weeks wos ... 
 
 looking for work or on layoff from o |ob' 
 
 H«b.. 
 5-14 wks. O 
 15-26 wks. C 
 
 27-39 wks. 
 40 wks. 
 
 38. You said ... worked about 
 
 (entry in item 34) weeks in 1973. 
 How many of the temoining C52 m 
 entry in item 34) weeks wOS ... 
 looking for work or on layoff 1.0. 
 a |ob? 
 
 (Mark weeks 
 and ask 39)^ 
 
 
 
 1 I 
 
 
 N ■ i. 
 (Sktp to 41) 
 
 39. Were the 
 
 Joking for work (or on layoff) oil in 
 ne stretch' 
 
 Yes -- 1 stretch . . 
 No -- 2 stretches. 
 
 No -- 3 + stretches 
 
 I (Sktp to 
 
 O \'" 
 
 37. What was the mom reason ... d.d not work in 1973? Was he 
 
 III or disabled ond unable to work 
 Taking care of home or family . . . 
 
 Going to school 
 
 Could not find work 
 
 In Armed Forces C 
 
 Retired O 
 
 Other O (Specify) ^ 
 
 (Sktp to .rem 45) 
 
 40. Did ... lose any full weeks of work 
 in 1973 because he wos on layoff 
 from a |ob or lost a |ob' 
 
 (Skip to 43) 
 
 41 INTERVIEWER CHECK ITEM 
 Are all weeks accounted for in 
 items 34 and 38' 
 
 >,.. 
 
 (Ski L, N ° < A,h42 > 
 
 to 43} 
 
 42- What was ... doing most of lh< 
 remaining weels .n 1973? 
 
 Was he ... 
 
 I 
 
 III or disabled ond 
 
 unable to work 
 
 Toking care of home or family 
 
 Going to school 
 
 In Armed Forces 
 
 Retired 
 
 Other (Specify) 
 
 / 
 
 (Ask item 43) 
 
 43. When ... 
 work full 
 
 working in 1973, d.d he usually 
 t or part time 7 
 
 44. What was ...'s longest |ob in 1973? _ 
 (Compare with entry in item 23) ^ 
 
 Some os item 23 O (Skip to item 45) 
 
 Dilferent from item 23 (Specify below 
 
 (or item 23 blank) and go to item 45) 
 
 44A. NAME OF EMPLOYER 
 
 44B. INDUSTRY 
 
 44C. OCCUPATION 
 
 44D. MAJOR ACTIVITIES 
 
 44E. CLASS OF WORKER 
 
 Privote P 
 
 Federal Gov't F 
 
 Stole Gov't. ' S 
 
 Local Gov't L 
 
 Self-emp. 
 
 , I Yes I 
 
 " C - I No (or farm,. SE 
 
 Without pay WP 
 
 (Ask item 45) 
 
 INCOME IN THE YEAR 1973 (true 
 LAST YEAR (1973) How much did . 
 
 i the space provided m items 45~48e and mark the appropriate dolled numbers.) 
 
 45- In wages or salary before any deductions 
 
 CPS-581 
 
 3 3 3 3 3 
 
 46- In net income from his own business or professional 47. In net income from hit 
 
 practice or partnership 7 
 
 CPS-581 
 
 MARK NET INCOME 
 
 Gross 
 income 
 
 G 
 
 : I 
 
 Q O 
 
 1 I I 
 
 3 3 
 
 business 
 expense 
 
 equols 
 
 Lost 3 5 
 
 ':■ 
 
 Money :. 
 
 G 
 
 ? z 
 
 ? 
 
 Net 
 income 
 
 49. Did . . . live in 
 this house four years 
 ago, that is on 
 March 1, 1970? 
 
 No (Ask 50) 
 
 / 
 
 CPS-581 O 
 
 MARK NET INCOME 
 
 Gross 
 
 income 
 
 ! : 
 
 o o o 
 
 III 
 
 Lost 5 5 5 5 
 money G G G G 
 
 Net 
 
 ncome 
 
 equols 
 $ 
 
 48. During 1973, did ... receive any money from 
 
 48a. -500101 Security 
 checks from the 
 U.S. Government- 
 
 Yes 
 
 '. 
 
 ■ Railroad Retiremen 
 checks from the 
 U.S. Government' 
 
 How much 
 altogether'' 
 
 O O 
 
 1 III 
 
 D D 3 
 
 6 6 6 
 ? ? ? 
 
 48b. •Estates, trusts, ot dividends' 
 
 Yes : No : 
 
 ■ Interest on sovings accounts 
 or bonds' 
 
 • Net rental income' 
 Yes : No 
 
 low much 
 altogether' 
 
 O O O O 
 
 1 I ill 
 
 3 3 
 
 3 3 3 
 
 3 3 5 5 S 
 
 Lost money G G G 6 6 
 
 O ? ? ? ? ? 
 
 ■Welfare payments of other 
 public assistance (aid to 
 families with dependent 
 children, old age assistance. 
 or aid to the blind Or totally 
 disabled/? 
 
 ' 7 
 
 •.. 
 
 (Determine specific AMOUNTS 
 and monlhs RECEIVED) 
 
 How much 
 altogether 
 
 CPS-581 
 
 O O O 
 
 1 III 
 
 6 6 6 
 ? 7 ? 
 
 48d. •Unemployment compen 
 
 Yes I No 
 
 ■ Workmen's compensotn 
 g| Yes 6 No 
 
 • Government employee | 
 Yes No 
 
 ■ Vetetan's payments' 
 
 Yes O No 
 
 How much 
 altogether' 
 
 CPS-581 
 
 O 
 
 1 I 
 
 i. "Private pensions or annuities' 
 Yes O No O 
 
 • Alimony' 
 
 Yes O No O 
 
 • Regular contributions from per - 
 sons not living in this household' 
 
 Yes O No O 
 
 • Anything else' 
 
 Yes J No O 
 
 How much 
 altogether' 
 
 CPS-581 
 
 c e 
 
 c c' c 
 
 O 
 
 3 3 
 
 3 3 3 
 
 
 °r "i- 
 
 °r °r °r 
 
 
 5 5 
 
 3 5 3 
 
 
 6 6 
 
 6 6 6 
 
 
 ? ? 
 
 ? ? ? 
 
 FOR OFFICE USE ONLY 
 INDUSTRY 
 
 ,;, , A 
 
 OCCUPATION 
 
 3 
 
 4 
 5 
 
 6 
 7 ■ 
 
 O O i i 
 
 . 
 
 / 
 
 8 O O 
 I 
 
 MIGRA- 
 TION 
 
 I I ! 
 
 : £ 6 '3. : 
 
 50. Where did . . . live on March 1, 1970' 
 a. Nome of State, foreign country, 
 
 U.S. possession, etc. 
 
 7 
 
 b. Name of county 
 
 ">' 
 
 51. Did . . . live inside the limits of a city, 
 
 town, village, etc' 
 
 (Enter name 
 Yes O of lown.flc.) No O (Go to 52) 
 
 52. What is . 
 (Shaw Flo. 
 
 1 German . 
 
 02 Italian . . 
 
 03 Irish.... 
 
 04 French. . 
 
 05 Polish., 
 6 Russian . 
 7 English | 
 
 08 Scottish J- 
 
 09 Welsh..] 
 Don't know 
 
 .'s origin or descent? ^_ 
 ,h Card or read list) ^ 
 
 O ilO Mexican-American. 
 
 O 111 Chicano 
 
 J ! 12-13 Mexican (Meccano) 
 
 C I 14 Puerto Ricon 
 
 O il5 Cuban 
 
 O i 16 Central or So. Amer. 
 
 I 17 Other Sponish 
 
 O ! 20-21 Negro or Black ... 
 
 | 30 Other (Specify below) 
 
 53a. Interviewer Check Item 
 
 Who reported on income for this person? 
 
 Self O ■ Other O 
 
 b. Type of Interview for this person 
 (One entry only) 
 
 Personal only ■' Telephone, regular C 
 
 Personal, Telephone, reguloA Q 
 
 telephone callback! telephone.collback/ 
 
 IF LAST PERSON, GO TO ITEM 54 
 
 137 
 
PLANNING THE FEDERAL STATISTICAL SYSTEM-A HISTORICAL PERSPECTIVE 
 
 Joseph Duncan, Deputy Associate Director for Statistical 
 Policy, Office of Management and Budget 
 
 "As individual economic and social problems have become evident or new social or 
 economic programs defined, the Congress has authorized specific statistical activities to 
 meet the needs of policymaking, program administration, and general analysis of social 
 well-being. For most of this century there has been a growing effort to initiate effective 
 planning and coordination of the many demands for statistical information associated 
 with both congressional concerns and executive requirements." 
 
 Joseph Duncan 
 
 Mr. Barabba: Our next speaker is Dr. Joseph Duncan. He is the 
 Deputy Associate Director of the Office of Management and 
 Budget responsible for statistical policy, and in that capacity he is 
 considered to be the chief statistician of our Federal statistical 
 system. This office, since Dr. Duncan came, has been asserting 
 itself as the coordinating center for the entire Federal statistical 
 system, although there's no central statistical office in the United 
 States as exists in many countries. There are advantages and 
 disadvantages of decentralization which Dr. Duncan will discuss, 
 as well as how his office and the rest of the statistical agencies 
 operate and make decisions. 
 
 Dr. Duncan: Throughout this series you have discussed a variety 
 of important statistical programs each of which makes significant 
 contributions to our Nation's total knowledge concerning eco- 
 nomic and social issues. Today I plan to review the development 
 of the Federal statistical system and outline current programs 
 which assure its continued development. 
 
 In undertaking this rather wide ranging assignment I, first, 
 would like to identify the principal conditions that created the 
 statistical system. Next, I plan to trace the historical develop- 
 ment of the Statistical Policy Division in the Office of Manage- 
 ment and Budget, which is responsible for planning and 
 coordinating this system. Finally, I will outline the nature of the 
 planning process presently being defined to respond to future 
 needs. 
 
 Factors Affecting System Development 
 
 If a statistical management expert were required to create a 
 Federal statistical system for the United States in 1974, assuming 
 no past statistical activity, it is doubtful that he would design the 
 sprawling monster that we cope with now. It has been estimated 
 that as many as 65 to 88 Federal agencies are authorized to 
 collect, compile, and disseminate statistical data. 1 Regardless of 
 the definition of "agency" and "statistics," it is clear there are 
 large numbers of separate and distinct organizations producing 
 statistics within the Federal Government. 
 
 How did this data collection begin? People sometimes 
 erroneously believe that government agencies established for 
 other purposes have launched statistical programs in an effort to 
 perpetuate or expand themselves. This is not so. Before I give you 
 
 some historical perspective on how the United States came to 
 have so many statistical cells within the agencies, let me give you 
 a modern example with which we are all familiar. 
 
 In the Federal Energy Administration Act of 1974, Section 
 13, entitled "Information-Gathering Power," gives the Admini- 
 stration broad powers to collect, assemble, evaluate, and analyze 
 energy information. This section also requires that all persons 
 involved in any phase of energy production or consumption make 
 records available to the administrator and respond to specific 
 questions asked by the administrator, who is empowered to 
 investigate and verify information and to subpoena records of a 
 statistical nature. This act is a prototype, in many respects, f 
 legislation requiring and empowering a Federal agency, newly 
 created in response to a critical national situation, to establish a 
 statistical program. 
 
 The first legislation calling for Federal statistical activity, 
 the Constitution of the United States, required an enumeration 
 of the population within 3 years after the first meeting of the 
 Congress and every 10 years thereafter. This was the origin of the 
 Census Bureau, with Thomas Jefferson as the first Director of the 
 Census. Alexander Hamilton, Secretary of the Treasury, issued 
 America's first statistical publication, reporting the population 
 and foreign commerce in 1790. Originally, the Census Bureau was 
 phased out after its decennial activity, but subsequent legislation 
 and responsibilities gradually closed the gap between censuses. 
 
 For example, the Director of the Census was "authorized 
 and directed to collect and publish, for every second year after 
 1927, statistics of manufacturing industries." 2 In 1929, it was 
 decreed that "a census of population, irrigation, drainage, distri- 
 bution, unemployment, and mines shall be taken by the Director 
 of the Census in the year 1 930 and every 1 years thereafter." 3 In 
 fact, one requirement was remarkably detailed: 
 
 The Director of the Census is authorized and directed to 
 collect and publish statistics concerning the number of 
 squares of red-cedar shingles produced in shingle manufac- 
 turing establishments in the United States; the quantity of 
 Canadian produced red-cedar shingles stored in U.S. ware- 
 houses and subsequently withdrawn therefrom; and the cur- 
 rent imports of red-cedar shingles from Canada. ...The statis- 
 tics as to the number of squares of shingles. ..shall relate to 
 each calendar month and shall be published as soon as 
 possible after the close of the month. 4 
 
 1 Stuart A. Rice, "American Government and Politics," The American 
 Political Science Review, Vol. XXXIV, No. 3, June 1940. 
 
 138 
 
 2 Act of June 18, 1929, C. 28, sec. 17,46 Stat. 25. 
 
 3 Act of June 18, 1929, C. 28, sec. 1,46 Stat. 21. 
 
 4 Act of May 25, 1937, C. 261 , sec. 1, 50 Stat. 204. 
 
PLANNING THE FEDERAL STATISTICAL SYSTEM-A HISTORICAL PERSPECTIVE 
 
 Statutory requirements for agriculture statistics were first 
 established in 1862. 5 In 1927, the Secretary of Agriculture was 
 directed to collect annual statistics and to publish his estimates of 
 grades and staple length of cotton of the then current crop. 6 In 
 1929, detailed statistics about tobacco were also authorized. 7 
 
 In 1888, the Bureau of Labor Statistics received its legisla- 
 tive mandate as follows: 
 
 ...to acquire and to diffuse among the people of the United 
 States useful information on subjects connected with labor 
 ...and especially upon its relation to capital, the hours of 
 labor, the earnings of laboring men and women, and the 
 means of promoting their materials, social, and intellectual, 
 and moral prosperity. 8 
 
 Federal statistical programs often originate in congressional 
 legislation. In fact, the economic and social historv of this 
 country could probably be traced around the statistical programs 
 that have emerged sequentially through legislation, particularly 
 when statistical activities associated with program administration 
 are also taken into account. A notable example is the collection 
 of statistics related to the various Social Security programs. 
 
 This process of generating a statistical system has given this 
 Nation the best statistical coverage in the world. But the effi- 
 ciency of this uncentralized system has been challenged period- 
 ically during the last several decades. Other countries such as 
 England and Canada meet their statistical needs most adequately 
 with centralized systems, and the concept has great appeal. 
 Inherent in a centralized system are the promises of coordination, 
 elimination of duplication, and economies of staff and equip- 
 ment. The subject has been carefully studied a number of times 
 by experts, but the impracticalities of centralizing Federal statis- 
 tical activities always overruled its conceptual neatness. Com- 
 plexities of legislation and program combine with vast geographic 
 spread and a large population to make it more feasible to do the 
 best we can with the system we have. 
 
 Another recurrent outcome of these studies of the Federal 
 statistical system is the recommendation for increased coordina- 
 tion. As background for our subsequent discussion, I should like 
 to trace the development of coordinating efforts leading up to the 
 establishment of the Statistical Policy Division in the Office of 
 Management and Budget. 
 
 Historical Trends in the Coordination of Statistics 
 
 As early as 1908, an interdepartmental statistical commit- 
 tee was established by Executive order to report on coordination 
 among statistical activities which were expanding to keep pace 
 with the growth of the Nation. 
 
 Wars give rise to the need for the production of statistics. 
 During World War I, the proliferation of statistical activities 
 associated with the war effort caused enough concern that in 
 1918 the War Industries Board established the Central Bureau of 
 Planning and Statistics. Some of the functions the Bureau under- 
 took were to improve statistical operations by developing uni- 
 form standards and definitions, cataloging government statistics, 
 and advising the agencies on statistical methods. Although this 
 
 s Act of May 5, 1862, C. 72, sec. 3, 12 Stat. 387. 
 6 Actof Mar. 3, 1927, C. 337, sec. 1,44 Stat. 1372. 
 7 Act of Jan. 14, 1929, C. 69, sec. 1,45 Stat. 1079. 
 "Act of June 13, 1888, C. 389, sec. 1,25 Stat. 182. 
 
 Bureau existed only until 1919 when its functions were trans- 
 ferred to the Bureau of Efficiency, some of the tasks it set for 
 itself read much like the present-day objectives of the Statistical 
 Policy Division of OMB. 
 
 The Bureau of Efficiency concerned itself with statistical 
 coordination from 1919 to 1922, and vainly recommended to 
 Congress centralization of nonadministrative statistical work in a 
 Federal Bureau of Statistics. This may have been the earliest of 
 several attempts to establish major centralization, efforts which 
 have consistently failed. 
 
 A Federal Statistics Board, established in 1931, was ineffec- 
 tive largely because of lack of authority and staff. By direction of 
 President Hoover in a Budget circular, the board was to study the 
 collection, compilation, and use of statistics and to recommend 
 economies and means for fuller utilization of statistics and statis- 
 tical personnel. 
 
 The successor agency, the Central Statistical Board, was the 
 earliest viable ancestor of today's Statistical Policy Division. Its 
 foundations, therefore, are of great interest as the precursor of 
 current coordination practices at the national level. The Board 
 was an independent agency in 1933 when the important statis- 
 tical issues of the times concerned the Government's efforts to 
 cope with the depression. The Board was initially limited to 
 reviewing and coordinating statistical activities for the National 
 Industrial Recovery Act. 
 
 In 1934, an Executive order, in addition to broadening the 
 scope of concern relative to the recovery program, directed the 
 Board to plan and promote the improvement, development, and 
 coordination of statistical services. It had authorization to investi- 
 gate any statistical service of a Federal agency, and certain non- 
 Federal statistical services. 
 
 The Chairman of the Board was Dr. Stuart A. Rice. When 
 the Board's functions were transferred to the Bureau of the 
 Budget, Dr. Rice was selected the Assistant Director of BOB for 
 Statistical Standards until his retirement at the end of 1954. 
 Many illustrious names of the 1930's were associated with the 
 Central Statistical Board and the Central Statistical Committee, 
 which served in an advisory and supervisory capacity. Frances 
 Perkins, then Secretary of Labor, was the Committee's Chairman 
 with Henry Morgenthau, Secretary of the Treasury; Henry 
 Wallace, Secretary of Agriculture; and Darnel C. Roper, Secretary 
 of Commerce, as Committee members. The Board, itself, had 
 leading government administrators of the era as members, 
 including Arthur Altmeyer, who led the Social Security Board 
 through its early days, Morris Copeland, who was the Executive 
 Secretary, and Federick Stephen, a leading international figure in 
 statistics. 
 
 I cite these names to demonstrate the great importance that 
 was increasingly attributed to good statistics during the Depres- 
 sion era, when urgent national policy was formulated on the basis 
 of fragmentary, inconsistent, and out-of-date numbers, only 
 awkwardly reflecting the economic state of the country. In con- 
 trast to the present system with its monthly and weekly eco- 
 nomic indicators, the Board in 1938 reported as follows: "The 
 Board has been urging for the past 3 years the quarterly collec- 
 tion. ..of a brief financial statement from a limited number of 
 corporations. ..to fill important gaps in business trends." 9 It was 
 modestly hoped that this procedure could be tried on an "experi- 
 mental basis, in the not too distant future." 
 
 Fourth Annual Report of the Central Statistical Board, June 30, 1938. 
 
 139 
 
PLANNING THE FEDERAL STATISTICAL SYSTEM-A HISTORICAL PERSPECTIVE 
 
 The Board concerned itself in the late 1930's with issues 
 such as the development of the Standard Industrial Classification. 
 The Board also reported "a strong and friendly interest in the 
 statistical activities of State governments." And in 1938, the 
 President asked the Board "to make recommendations designed 
 to lighten the burden (on business and the public) of reports." 
 
 As always, Congress was increasing reporting requirements 
 while the executive branch of government was trying to hold 
 them down. For example, in May 1938, Congress passed an act 
 broadening the scope of statistics required of processors of pea- 
 nuts and peanut oil. However, since some of the required data 
 were already being collected by the Census Bureau, the Board was 
 successful in eliminating the duplication in the statistics of pea- 
 nuts. 
 
 By 1938, it was estimated by the Board that 88 Federal 
 agencies collected reports from the public. At that time, about 
 three-quarters of Federal statistical work was "administrative" 
 and connected with governmental services and regulatory activi- 
 ties, like Social Security, relief, transportation, and income tax 
 statistics, much of which had useful general byproducts. 
 
 In 1940, Stuart Rice reported the following: 
 
 The Federal statistical system has many of the character- 
 istics of the national statistical system of which it is a part 
 and of the social economy which it supplements. It is made 
 up of a large number of semicompetitive, semimonopolistic, 
 semi-integrated agencies, without very effective central 
 control. This is true also of the administrative agencies to 
 which the statistical agencies are subjoined. 10 
 
 However, in spite of this proliferation of statistical responsi- 
 bility, the Board recommended official acknowledgement that 
 various Federal statistical services be attached to those agencies 
 with responsibilities relating to the subjects of the reports. Also, 
 because of the decentralization, a statistical coordinating agency 
 with adequate powers was essential. Rice concluded. 
 
 Historically, statistics originated with the agency programs 
 and services. General-purpose statistics necessary for broad 
 informational and policymaking reasons were concentrated then, 
 as they are now, in relatively few agencies. 
 
 The Central Statistical Board was transferred to the Bureau 
 of the Budget by the Reorganization Act of 1939 which also 
 transferred BOB from the Treasury Department to the Executive 
 Office of the President. The Board's functions remained essential- 
 ly unchanged until the Federal Reports Act of 1942" greatly 
 enlarged the division's scope of work. That and the later amended 
 Budget and Accounting Act of 1921, 12 were the statutory bases 
 for Federal statistical coordination. Prior to that time it was 
 recognized that coordination was needed, but no real statutory 
 basis for that coordination existed. This legislation recommended 
 minimizing respondent burden, investigating data-gathering meth- 
 ods, and coordinating the "information-collecting" services of 
 government. The laws are specific as to agencies excluded, release 
 of information, and necessary actions. 
 
 Student: Is your only function the elimination of duplication? 
 What happens when forms come across your desk and you think 
 it's absolute idiocy to approve them even if duplication is mini- 
 mal? 
 
 Dr. Duncan: We have several criteria in addition to duplication. 
 For example, we look at survey methodology, and we also have 
 some informal guidelines with regard to individual privacy; e.g. 
 "are these reasonable questions to be asked?" Our main approach 
 right now, however, is to determine the basis of the suggested 
 data collection. If there is insufficient reason for seeking the infor- 
 mation we will disapprove it even if no one else is collecting it. 
 
 The 1950 Budget and Accounting Act, Section 103, 
 instructs the Director of the Bureau of the Budget to develop and 
 administer programs for the improved gathering, compiling, 
 analyzing, publishing, and disseminating of statistical information 
 for any purpose by the various agencies in the executive branch of 
 the Government. This responsibility and authority were delegated 
 to the Assistant Director for Statistical Standards, now entitled 
 Deputy Associate Director for Statistical Policy. 
 
 During World War II, the responsibilities of the office were 
 compounded by the necessity of approving questionnaires associ- 
 ated with the war effort. New agencies like the Office of Price 
 Administration and the War Production Board were brought 
 under the control of the Budget Bureau to keep necessary report- 
 ing by business and industry to a reasonable minimum. In 1943, 
 the first year after the enactment of the reporting law, about 
 7,500 questionnaires were submitted of which about 12 percent 
 were disapproved. 
 
 After the war, the Division's forms clearance responsibilities 
 diminished to about 2,500 to 3,000 requests in the mid-1 950's— 
 a level which has remained fairly constant— and coordination and 
 integration efforts were increased. The Division of Statistical 
 Standards, changed to the Office of Statistical Standards in 1952, 
 encouraged the agencies to establish internal clearance units to 
 sift out unacceptable reporting proposals. Also, the office was 
 receiving assistance from numerous interdepartmental commit- 
 tees. 13 In the fall of 1951, the American Statistical Association 
 organized an Advisory Committee on Statistical Policy whose 
 membership was then composed of past presidents of the Associa- 
 tion. It is interesting to note that a leading topic of the first 
 meeting of the ASA Advisory Committee, confidentiality, is still 
 a major agenda item for that committee. 
 
 In the early 1950's, considerable emphasis was also given to 
 the development of documents defining statistical standards, like 
 Standards for the Publication of Statistical Data, Standard 
 Commodity Classification, and Standard Industrial Classification. 
 These standards promoted comparability and helped to control 
 the quality of statistical collection and tabulation. 
 
 Additional Roles for the Office of Statistical Standards 
 
 During the 1950's, the role of the Office of Statistical 
 Standards expanded. In addition to its functions concerned with 
 reviewing agency data-collection proposals and setting standards, 
 the office also reviewed the adequacy of statistical series and 
 planned for new series to fill statistical gaps. The office partici- 
 pated in international activities preparing position papers on sta- 
 tistical matters, selecting U.S. data for United Nations publica- 
 tions, and representing the U.S. on statistical advisory bodies to 
 international organizations. 
 
 In 1955, Dr. Raymond T. Bowman became the Assistant 
 Director for Statistical Standards of the Bureau of the Budget, 
 
 10 Stuart Rice, op. cit. 
 
 11 Public Law 831, December 24, 1942. 
 
 12 Public Law 13, June 10, 1921; Public Law 784, September 12, 1950. 
 
 13 "The Development of Statistical Coordination," by Clem C. 
 Linnenberg, Jr., The American Statistician, June-July 1949. 
 
 140 
 
PLANNING THE FEDERAL STATISTICAL SYSTEM-A HISTORICAL PERSPECTIVE 
 
 replacing Dr. Rice. In addition to directing the work of the Office 
 of Statistical Standards, he wrote extensively on statistical 
 developments. In an address before the Washington Statistical 
 Society, in January 1957, Bowman noted that although the sta- 
 tistical program of the U.S. was outstanding, it was not good 
 enough for the demands made upon it— a statement likely to be 
 forever applicable. Recognizing the impact of computers on the 
 capability for producing fast, detailed information, he urged 
 special attention to the quality of statistics, particularly referring 
 to the work being done by the Census Bureau and the Bureau of 
 Labor Statistics in the field of response error. Sampling had been 
 introduced in national surveys at the end of the 1930's, and 
 techniques had steadily improved during the 1940's and 1950's 
 while nonsampling error still required great research and experi- 
 mental efforts. 
 
 Quality was emphasized by the Office of Statistical Stand- 
 ards (OSS), as were detailed improvements in economic statistics. 
 The Budget Bureau in 1956 contracted with the National Bureau 
 of Economic Research for an independent appraisal of the 
 national income and product accounts. It encouraged the greater 
 development of a joint Security Exchange Commission and 
 Federal Trade Commission (SEC-FTC) quarterly report for manu- 
 facturing, trade and mining corporations. 
 
 The Office of Statistical Standards arranged for the Federal 
 Reserve System to take leadership in developing and interpreting 
 personal savings data. Also, greater use of administrative records, 
 such as IRS records and BOASI (Bureau of Old-Age and Survivors 
 Insurance) records, was encouraged as a relatively inexpensive 
 source of statistical information. OSS efforts at integration and 
 coordination were fully focused on improving economic series 
 such as employment, unemployment, and construction statistics, 
 and manufacturers' sales, inventories and new orders. In the 
 1950's, as before, social statistics were not greatly emphasized in 
 the Office of Statistical Standards. 
 
 For the next few years, through the mid-1960's, the 
 Federal statistical system grew in complexity. The agencies were 
 having problems of internal coordination as their statistical 
 programs grew in response to increased demand for data. The 
 Census Bureau was increasing its data-collection and processing 
 services for other agencies on a reimbursable basis, and outside 
 contracting grew as a resource for agencies incapable of their 
 own. 
 
 Although in 1964 the OSS staff was reduced to less than 40 
 persons, the activities of the office expanded in several directions. 
 Since the central coordinating agency, the Office of Statistical 
 Standards, was formally entrenched in a good legislative base with 
 three decades of experience behind it, the efforts of the office 
 often took on a shape determined by the specific backgrounds, 
 abilities and leadership qualities of the personnel. The Federal 
 statistical agencies also tended toward shaping their programs on 
 the basis of a focal agency concept. That is, although several 
 agencies might be engaged in one way or another in a specific 
 statistical activity, one agency was generally recognized as the 
 lead agency, exercising its coordinating efforts with the support 
 of the Office of Statistical Standards. Hence, this aspect of 
 coordination was viewed as based on an interacting principle. 
 
 OSS continued to regard several other devices as coordi- 
 nating instruments, such as the review of agency statistical bud- 
 gets, data requests, interagency committees and task forces, non- 
 governmental advisory groups, the use of technical consultants, 
 promulgation of standards, and the issuance of a number of publi- 
 cations relative to the operations of the Federal statistical system. 
 
 Toward the end of the 1950's when inflationary pressures 
 following World War II were building up to serious proportions, 
 greater emphasis was applied to work in price statistics, especially 
 in connection with the CPI (Consumer Price Index) and the WPI 
 (Wholesale Price Index). Another reflection of the economic 
 stress of the time was the effort being made and encouraged by 
 OSS to improve construction and retail sales statistics. An in- 
 teresting omen, in 1965, OSS chaired a Petroleum Study Group 
 directed toward developing better measures of reserves, produc- 
 tive capacity, wells, and revenues. 
 
 Furthermore, social and demographic statistics were being 
 given much more recognition as new social action programs were 
 developed. The Census Bureau was gearing up for its 1970 Census 
 of Population and Housing and OSS took its usual role in assisting 
 with the coordination of Federal agencies' needs for census data. 
 More attention was given to the relationship of Federal, State, and 
 local government statistical activities. Dr. Bowman was most 
 interested in exploring the feasibility of a Federal statistical data 
 center. However, it never enjoyed much popularity and eventual- 
 ly was completely rejected because of serious questions concern- 
 ing confidentiality Confidentiality issues continued to be 
 important as they always were in the coordinating efforts of the 
 Office of Statistical Standards. 
 
 Recent Developments in the Statistical Policy Division 
 
 In mid-1969 Julius Shiskin became the Assistant Director 
 for Statistical Policy in the Bureau of the Budget. Mr. Shiskin 
 viewed the office as having a broader function than that of setting 
 statistical standards and proposed the new name of "Office of 
 Statistical Policy." At the time Mr. Shiskin was appointed, 
 Robert Mayo was the Director of BOB, soon to be succeeded by 
 Dr. George Shultz who performed a dual role. Dr. Shultz was not 
 only the Director of OMB (BOB's successor) but also one of 
 President Nixon's principal economic advisers. Mr. Shiskin's back- 
 ground in the Bureau of the Census put him in a unique position 
 to be of assistance to the Director and, hence, the President, in 
 obtaining the most up-to-date economic information for use in 
 national planning and policymaking. Much of the activity of the 
 Office of Statistical Policy was directed toward improving 
 economic statistics, chiefly the principal economic indicators. 
 
 President Nixon instructed the Director of the Bureau of 
 the Budget to improve the timeliness of the principal economic 
 indicators, many of which were disseminated too late after their 
 reference periods to be of prompt value in policy formulation. 
 Also, the President required a time separation between the publi- 
 cation of statistics and interpretations issued by policymaking 
 officials. 
 
 These requirements were soon translated by the Office of 
 Statistical Policy into memorandums to the agencies. The first of 
 these required an advance schedule of publication dates for 
 principal economic series be reported in the Statistical Reporter. 
 The agencies were also required to attempt to reduce processing 
 time between the close of a reference period and the publication 
 of statistical information. It was also made a requirement that 
 there be a separation of at least 1 hour between the issuance of 
 statistical data and policy commentary by an appropriate official. 
 
 Another troublesome aspect of economic statistics was the 
 frequency and magnitude of revisions made in the economic 
 series. GNP figures were particularly affected by revisions so early 
 estimates of GNP sometimes proved to be quite misleading. In 
 
 141 
 
PLANNING THE FEDERAL STATISTICAL SYSTEM-A HISTORICAL PERSPECTIVE 
 
 some cases, such as retail sales, greater budgetary provision was 
 made to accommodate increased efforts toward improving the 
 accuracy and timeliness of the indicators. 
 
 Major emphasis in the office was also given to the develop- 
 ment of a pioneer publication. Social Indicators. While the office 
 recognized the importance of this work, it was able to devote 
 only one full-time position to its accomplishment. However, 
 several persons from the Census Bureau were instrumental in the 
 development of this major project. The concept of a consolidated 
 statistical budget was emphasized and the interactions with the 
 agencies in developing budgets were increased. 
 
 When President Nixon introduced his proposals for the 
 consolidation of the departments of Government into four major 
 departments, the Office of Statistical Policy proposed corre- 
 sponding consolidation of statistical activities to parallel the 
 departmental configuration. While the consolidation of govern- 
 ment departments did not receive Congressional support, it did 
 appear useful to continue to try to bring together the statistical 
 activities of Government in a more economic framework. 
 
 At the Department of Commerce the Social and Economic 
 Statistics Administration (SESA) was established to include the 
 Census Bureau and the Bureau of Economic Analysis. It was 
 hoped that a similar pattern could be developed among other 
 agencies, concentrating data-collection activities and analytic 
 capabilities to serve several organizational units. However, the 
 complexities of achieving centralized statistical agencies, even 
 within the major departments, have forestalled any real move- 
 ment toward consolidation. 
 
 In the meantime, at the beginning of President Nixon's 
 second term the scope of activities of the Budget Bureau were 
 enlarged to include management functions and the office was 
 renamed the Office of Management and Budget; Roy Ash is its 
 present Director. This reorganization also resulted in changing the 
 Office of Statistical Policy into the Division of Statistical Policy, 
 within the management and operations part of OMB. 
 
 Statistical Planning and Coordination in the Mid-1970's 
 
 During the earlier discussion I tried to demonstrate the 
 episodic character of the Federal statistical system. As individual 
 economic and social problems have become evident or new social 
 or economic programs defined, the Congress has authorized 
 specific statistical activities to meet the needs of policymaking, 
 program administration and general analysis of the social well- 
 being. For most of this century there has been a growing effort to 
 initiate effective planning and coordination of the many demands 
 for statistical information associated with both congressional con- 
 cerns and executive requirements. Presently, the major com- 
 ponents of the U.S. Federal statistical system include the follow- 
 ing: 
 
 1. General-Purpose Statistical Agencies 
 
 A. Bureau of the Census (U.S. Department of Commerce) 
 
 B. Bureau of Labor Statistics (Department of Labor) 
 
 C. Statistical Reporting Service (Department of Agri- 
 culture) 
 
 D. National Center for Health Statistics (Department of 
 Health, Education, and Welfare) 
 
 E. National Center for Educational Statistics (Department 
 of Health, Education, and Welfare) 
 
 2. Analytical and Research Agencies 
 
 A. Council of Economic Advisers (Executive Office of the 
 President) 
 
 B. Economic Research Service (Department of Agricul- 
 ture) 
 
 C. Bureau of Economic Analysis (U.S. Department of 
 Commerce) 
 
 D. Board of Governors of the Federal Reserve System 
 
 E. Office of Policy Evaluation and Research, Manpower 
 Administration (Department of Labor) 
 
 F. Consumer and Food Economics Institute, Agricultural 
 Research Service (Department of Agriculture) 
 
 G. Bureau of Domestic Commerce (Department of Com- 
 merce) 
 
 H. Bureau of International Commerce (Department of 
 
 Commerce) 
 I. Bureau of Mines (Department of the Interior) 
 J. National Energy Information Center (Federal Energy 
 
 Administration) 
 
 In broad terms the organization of the Federal statistical 
 system might be characterized by identifying the Statistical 
 Policy Division of the Office of Management and Budget as the 
 central office responsible for achieving an integrated and func- 
 tional system of accurate governmental statistics, improving 
 quality and reducing duplication in the data-collection and 
 statistical functions. 
 
 The Federal system of Government including its statistical 
 agencies involves many pluralistic approaches to various functions 
 including division of labor among State and Federal govern- 
 ments. The general-purpose statistical agencies collect, compile, 
 and publish data for general use in specified fields. The analytic 
 and research agencies are major users of data collected by these as 
 well as administrative and regulatory agencies and private sources. 
 The analytical and research agencies do in some cases undertake 
 direct collection of statistics as part of their activities. Finally, it 
 should be noted that most of the major administrative, regulatory, 
 and defense agencies collect statistical information connected 
 with their administrative and operating responsibilities, such as 
 the collection of taxes or regulation of public utilities. A few even 
 have specialized statistical responsibilities such as the data collec- 
 tion on scientific activities associated with the National Science 
 Foundation. 
 
 To complete the picture it should be noted that the private 
 sector relies heavily upon Federal data collection and analysis in 
 its daily decisionmaking. For example, marketing research depart- 
 ments of large corporations intensively utilize current economic 
 data, the decennial census results, and published regulatory infor- 
 mation to determine market strategy. Many individual citizens 
 evaluate their own situations in terms of general demographic 
 profiles concerning basic patterns, such as trends in cost of living 
 and wage settlements. State and local governments rely heavily 
 upon Federal statistics in an effort to formulate effective policies 
 within their areas of jurisdiction. 
 
 In summary, it is evident that the Federal statistical system 
 is a collection of relatively independent activities designed to 
 meet relatively specific needs associated with a wide range of user 
 groups including Congress, the executive branch, and the private 
 sector. 
 
 For the past 20 years there has been a specific effort to 
 reduce the burden of statistical data collection on individual 
 respondents. As mentioned earlier, the primary policy device for 
 
 142 
 
PLANNING THE FEDERAL STATISTICAL SYSTEM-A HISTORICAL PERSPECTIVE 
 
 accomplishing this objective is the Federal Reports Act of 1942 
 which is administered by the Statistical Policy Division. To evalu- 
 ate the many demands for data collection an overall set of priori- 
 ties for the statistical system is essential, as is measurement of 
 efforts to coordinate programs and data of various agencies. 
 
 In the past, an important device for enhancing this coordi- 
 nation and planning has been a variety of interagency commit- 
 tees, many of which are formally chaired by the Statistical Policy 
 Division. (See appendix.) Many other committees are appointed 
 for specific needs on an ad hoc basis. The Statistical Policy 
 Division is also advised by committees outside government. These 
 help SPD maintain wide perspective on national statistical issues. 
 
 A second technique for achieving planning and coordina- 
 tion is through the budget process in which the individual 
 agencies present their plans, programs, and request for resources, 
 both dollars and manpower. In each budget review all agencies are 
 questioned concerning statistical plans and programs, and 
 decisions are made concerning the overall direction of the Federal 
 statistical program. As final statistical budget decisions are de- 
 veloped they are included in the budget proposed to the Con- 
 gress. 
 
 It should be noted, however, that there are a number of 
 factors which influence the final outcome of the statistical pro- 
 gram not represented by the planning and budget decisions. These 
 include the following: 
 
 1. Budget proposals of statistical agencies are typically 
 screened within their departments priot to submission. 
 Consequently, needs perceived by statisticians in some 
 agencies may not receive full attention. 
 
 2. The final statistical budget decisions must be satisfacto- 
 rily integrated to overall departmental decisions. Statis- 
 tical programs which have widespread support some- 
 times do not survive the interdepartmental priority 
 process even if generally supported. It should be noted 
 that, in program agencies, provisions for statistical 
 budgets are not always clearly identified as such and are 
 regarded as part of the overall program. In that case, the 
 statistical provisions generally have the same budget fate 
 as the programs to which they are attached. 
 
 3. Proposals of the executive branch, as presented in the 
 President's budget message, may not be enacted as 
 proposed. The Congress, during appropriation hearings 
 and voting, frequently adjusts statistical programs by 
 eliminating some and requesting others not previously 
 included, 
 
 4. Finally, after programs are established, the evolution of 
 programs within the basic budgets of agencies frequently 
 develops on an incremental basis which permits many 
 adjustments and revisions which were not contemplated 
 in the initial planning. Typically, these take the form of 
 program improvements and result in more efficient data 
 collection or analysis. However, occasionally there are 
 instances where an agency decision to eliminate a parti- 
 cular survey or item of data has implications for users 
 who are not participants in the initial decision to revise a 
 survey. 
 
 On an overall basis this complex process has yielded a sta- 
 tistical system which is unsurpassed in the world. This certainly 
 speaks well for past decisionmaking and for the talents of the 
 statisticians who work throughout the Federal Government. 
 
 An Evolving Planning System 
 
 Despite the major accomplishments which have been noted 
 and the long-term effort at coordination it is evident that several 
 major problems exist in 1974. Simply stated these include- 
 
 1. Large-scale programs of data collection relative to social 
 problems have grown rapidly and in a relatively unco- 
 ordinated fashion, as a result of the social program 
 initiatives of the mid-1960's. In the absence of a social 
 accounting framework equivalent to the national eco- 
 nomic accounts it is particularly difficult to establish 
 conclusive priorities among these expensive and de- 
 manding social statistical programs. 
 
 2. The inflationary pressures of the past 2 years have 
 created exceptional demands on economic statistics and 
 revealed several methodological weaknesses in concep- 
 tual programs designed to deal primarily with relatively 
 price-stable environments. 
 
 3. Pressures to increase confidentiality of individual rec- 
 ords are in direct conflict with the trend during the past 
 20 years to use administrative records for statistical pur- 
 poses. Techniques must be identified for meeting the 
 dual objectives of protecting individual records and at 
 the same time providing mechanisms for interchange for 
 statistical purposes which will not affect the rights and 
 privileges of the individual and which will assure con- 
 sistent data. 
 
 In order to meet these problems more effectively the Sta- 
 tistical Policy Division is initiating a more structured planning and 
 coordination program in addition to the normal budget and 
 forms-clearance functions which have been so effective in the 
 past. 14 The new process focuses on developing a longer time 
 perspective on statistical programs by identifying major long-term 
 data programs and assuring even broader participation in priority 
 setting. 
 
 The details of this process are evolving but the results will 
 be a product of decisions made through the Interagency Commit- 
 tee on Statistical Policy and Programs which includes the heads of 
 all major statistical agencies. 
 
 Let me now turn to current activities of the Statistical 
 Policy Division as they relate to developing and leading a planning 
 process for the Federal Statistical System. In order to systematize 
 the discussion I would like to refer to the flow chart which 
 identifies some of the elements of the planning process. 
 
 Description of Steps in Statistical Planning Process 
 
 The flow chart illustrates the iterative nature of the 
 statistical planning process. Consequently, the following outline 
 is descriptive and does not represent a sequential statement: 
 
 1 . Formulate ultimate objective -the ultimate objectives of 
 the statistical system are derived from government policymakers, 
 congressional priorities, and private decision requirements. 
 
 14 Statistical Reporter, "Developing Better Long Range Plans for 
 Federal Statistics," October 1974, pp. 49-54. Published subsequent to 
 presentation. 
 
 143 
 
PLANNING THE FEDERAL STATISTICAL SYSTEM-A HISTORICAL PERSPECTIVE 
 
 2. Assess systemwide needs— the needs of system are de- 
 fined by comparing ultimate objectives with the existing statis- 
 tical programs. Needs reflect gaps in the present system. 
 
 3. Identify constraints and resources— the constraints on 
 the total Federal statistical system are the budget and personnel. 
 A long range plan must address the full cost of statistical pro- 
 grams including all stages of design, collection, processing, dis- 
 semination, and analyisis. The personnel available must be 
 balanced in terms of providing training and staff development of 
 new staff and utilization of existing staff. 
 
 4. Develop systemwide objectives— these objectives may be 
 related to the total statistical system or to a subsystem such as 
 the census operation. 
 
 5. Develop program structure— given the systemwide objec- 
 tives, the program structure is developed by examining alternative 
 methodologies and selecting the most appropriate one in relation 
 to available resources. 
 
 6. Perform program analysis— the selection of the most 
 appropriate approach also results in internal improvement of the 
 program structure. 
 
 7. Develop program budget— the program budget is the 
 result of relating available resources with alternative methodolo- 
 gies to yield the optimum allocation of resources. 
 
 8. Allocate resources— because available resources are allo- 
 cated among several programs there is a requirement to reexamine 
 program priorities before the final allocation of available re- 
 sources. 
 
 9. Operate the system— the implementation of the meth- 
 odology may require us to augment or reduce allocated resources 
 as experience determines the utility of the budget decision. 
 
 10. Evaluate programs— as the system functions the evalua- 
 tion of the final product may require modifications in operation 
 of the system or result in the definition of major gaps in statis- 
 tical information in relation to the overall needs. 
 
 Steps in the Statistical Planning Process 
 
 (D 
 
 
 Formulate 
 Ultimate 
 Objective 
 
 
 (2) 
 
 
 (4) 
 
 ' 
 
 
 (3) 
 
 Assess 
 
 Systemwide 
 
 Needs 
 
 
 Develop 
 
 Systemwide 
 
 Objectives 
 
 
 Identify 
 
 Constraints and 
 
 Resources 
 
 
 
 t 
 
 . 
 
 (6) 
 
 
 (5) 
 
 ' 
 
 
 
 
 
 
 Perform 
 Program 
 Analysis 
 
 
 Develop 
 Program 
 Structure 
 
 
 
 
 
 
 
 
 
 
 
 /yi "Top Down" 
 
 
 
 
 "Bottom Up" 
 
 Develop 
 Program 
 Budget 
 
 < 
 
 
 
 
 
 
 
 
 
 
 
 (8) 
 
 r 
 
 
 
 
 
 Allocate 
 Resources 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 (9) 
 
 r 
 
 
 
 
 Operate 
 The System 
 
 
 
 
 
 
 
 
 
 
 (10) 
 
 r 
 
 
 
 
 
 Evaluate 
 Programs 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 144 
 
PLANNING THE FEDERAL STATISTICAL SYSTEM-A HISTORICAL PERSPECTIVE 
 
 In reality much of the planning of statistical programs 
 begins in box 5. Individual agencies, based upon legislative 
 requirements, establish a program of statistics. The cost of the 
 program is estimated (box 7) and, within individual departments, 
 the budget for statistics competes with other program demands. 
 
 Also, in the present system as the statistical programs 
 operate (box 9) there is evaluation by the Statistical Policy 
 Division and by users. This evaluation feeds back on the 
 operation of the system. 
 
 A significant part of program innovation actually occurs 
 through the "bottom-up" process where statisticians in agencies, 
 based upon analysis of statistical programs and their deficiencies, 
 develop program budgets for improved utilization of resources to 
 achieve the overall statistical needs. The other boxes on this chart 
 are not totally inactive, however they are not systematically em- 
 ployed. Our objective in developing the statistical system plan- 
 ning process is to activate remaining boxes and develop more ef- 
 fective processes for priority setting in allocation of resources. 
 
 It is our view that this relatively straightforward planning 
 process will be effective only if it is characterized by active 
 participation. For example, the definition of needs involves many 
 kinds of users of Federal statistics such as policymakers within 
 the government, analysts in the various agencies, and the private 
 sector including business, academia, and individual citizens. 
 
 A fundamental responsibility, therefore, is to assure that 
 the planning process provides adequate opportunity for partici- 
 pation and involvement of both Federal statisticians and various 
 user representatives. Additionally, given the vast scope of this 
 enterprise, when one considers the range of Federal statistical 
 activities, it is evident that the primary role of the Statistical 
 Policy Division must be catalytic, with the substantive effort 
 largely accomplished through the major statistical agencies. The 
 primary responsibility of the Statistical Policy Division is that of 
 the integration of the perspectives of the various agencies and a 
 balancing of needs and demands for staff and budget across the 
 governmental structure. 
 
 The tradition of a central focal point for statistical planning 
 and coordination has been well established; and it works. 
 Obviously, there has to be some kind of effort to avoid 
 fragmentation of statistical activities. And, I think we have moved 
 a long way in the proper direction of developing adequate 
 collection capabilities in major agencies such as the Census 
 Bureau, assuring major statistical requirements are well covered 
 professionally. 
 
 However, I do see an important need to do a more 
 systematic job of defining the needs for information within the 
 agencies and among the many decisionmakers, both within 
 governments and the private sector. For example, in order to 
 better define these needs, beginning this month, we are to hold 
 conferences with State and local governments to determine 
 requirements from their prespectives. Having defined and ana- 
 lyzed various needs, we shall work with the statistical agencies in 
 establishing priorities in terms of their work and programs. 
 
 The Office of Management and Budget is concerned, on a 
 yearly basis, with the priority question of applying limited 
 resources to proposed statistical programs. The flow chart was an 
 attempt to illustrate the procedures for identifying the various 
 participants' activities and distributing funds from a total 
 resource of $300 million to $400 million. Unfortunately, $300 
 million in a $300 billion budget will not receive the prime 
 attention of all of the budget examiners. I view an important part 
 
 of my job as providing a rational and professional judgement con- 
 cerning the statistical programs we recommend. 
 
 Mr. Barabba: Let me just amplify a problem that further 
 compounds this situation. For the Bureau's reimbursable activ- 
 ities, we receive the dollars to conduct these activities from other 
 agencies or from the public sector. But this does not mean that 
 we are allowed a corresponding allocation for staff. So, we have 
 to ask Dr. Duncan to negotiate a waiver, if necessary, for 
 additional personnel. 
 
 Another example results from the energy crisis. We are 
 allowed a specified number of miles of reimbursable travel. Thus 
 we may have the personnel and the dollars to conduct a survey 
 but need Dr. Duncan's assistance to obtain permission for 
 additional reimbursable travel. 
 
 Dr. Duncan: Now I would like to believe that what I am about to 
 say is not true but I haven't had any evidence to the contrary. 
 Basically the system stops at box 9 of the chart. For example, 
 from an efficiency point of view there is only a limited amount of 
 internal evaluation of data collection. The system certainly does 
 not have a complete evaluation of systemwide needs going back 
 to systemwide priorities. There is a small amount of feedback 
 within agencies, such as program evaluation of procedures, 
 techniques, and utility of surveys. 
 
 I'm hoping to establish the mechanics of this chart as a real 
 process. That's not going to happen in a very short time. It's 
 going to take a number of years, at least. 
 
 We've mapped out a program which has some pilot 
 activities for the fiscal 1976 budget season which starts in 
 September. The overall planning process is really gearing up for 
 fiscal year 1977. Let me tell you about some of the elements 
 within that planning process. 
 
 First, during the fiscal 1976 budget cycle, we are going to 
 undertake a partial evaluation of programs in wage and price 
 statistics within the Bureau of Labor Statistics. Consideration is 
 made of program purpose, authorization, techniques, cost and 
 appropriateness. 
 
 Second, we, through our interagency committees, have 
 established programs to determine existing data needs. For 
 example, a number of departments need better information on 
 family income, expenditure, and wealth. Some agencies would 
 like to know how much income is spent on various kinds of 
 health care, and so they do a specialized survey. It could be more 
 appropriate to combine the health expenditure survey with a 
 survey of total consumer expenditures to provide a better 
 framework. 
 
 Also, we are presently holding meetings around the country 
 to learn the statistical needs of State governments. We know the 
 legislative demands for revenue sharing and related activities, but 
 we would like to develop programs to help the States meet both 
 their own and Federal needs. 
 
 We have had, for many years, active outside advisory 
 committees. For example, the Federal Statistics Users' Confer- 
 ence represents a formal structure complete with staff that is 
 available to advise the whole Federal statistical system on behalf 
 of business, labor, nonprofit institutions, and associations. The 
 Business Advisory Council on Federal Reports has business 
 representatives who help assess the reporting burden implied by 
 various survey proposals submitted to OMB. 
 
 Student: From what specific areas do you draw these people? 
 
 145 
 
PLANNING THE FEDERAL STATISTICAL SYSTEM-A HISTORICAL PERSPECTIVE 
 
 Dr. Duncan: They're drawn from large businesses. Typically, 
 they represent comptrollers' offices of very large companies- 
 people who actually fill out survey forms. -A number of 
 associations are represented, including several small business 
 associations. The Chamber of Commerce also represents small 
 businesses . It's hard to get small businessmen to devote a lot of 
 time to advisory committees, but we try to give attention to their 
 specific concerns. We also have professional groups that advise us. 
 
 Today we are asking for input about priorities and concepts 
 of statistical systems. A little later, as we begin to see new 
 program needs, we will have both internal and external groups 
 discuss individual proposals for programs. 
 
 Under the Budget Reform Act of 1974 an organization 
 equivalent to OMB is being established in the Congress. I don't 
 know if Congress intends to include a division for statistical 
 policy, but there certainly will be provision for defining the 
 informational needs of Congress. Recently, congressional com- 
 mittees have been fairly explicit as to what they would like to 
 have, but we're hoping for a broader statement of congressional 
 information needs. 
 
 My main point is that the definitions of the ultimate and 
 systemwide objectives have to involve statistical systems of 
 agencies and all of the major users and this is a long process. A 
 number of activities are beginning, in a piecemeal way, to identify 
 these goals in a fashion which can contribute to a more rational 
 planning system. 
 
 Let me just mention some other items which are central to 
 the system the flow chart describes. Within the budget process, I 
 hope we can move in the direction of a distinct statistical budget; 
 an identification of constraints and resources in terms of both 
 people and money for the total statistical community. 
 
 I feel strongly about the people side of the system. We 
 don't really have a good measure today of the skills and 
 capabilities of statisticians in our Government. We don't have the 
 kind of system that exists in some countries. Great Britain, for 
 example, has a Civil Service corps called Statisticians. In their 
 fifth year of service the Chief Statistician in Britain interviews 
 every statistician about his or her career plans. He then follows 
 through working on personnel exchange and training of the more 
 promising statisticians. I sense a need to make the job of Federal 
 statistician an enticing one with great opportunities for the 
 individual, in order to bring in needed new talent to help solve 
 the great statistical problems facing us. Mr. Barabba has taken this 
 approach within his agency and we have other agency heads that 
 do so as well. 
 
 Student: Is this selection of statisticians controlled by the Civil 
 Service? 
 
 Dr. Duncan: No, the Civil Service does not operate on a 
 job title basis. I think there is a need to identify the career goals 
 of young people in government. If we are going to talk about 
 improved Federal-State cooperative programs, we will also have 
 to do a great deal to influence the development of statistical 
 capability, professionally, at the State level. 
 
 Student: Isn't there a general mission for the statistical bureaus 
 in each of the executive departments to collect the data upon 
 which their programs are implemented and evaluated? 
 
 Dr. Duncan: There are a great variety of purposes for the 
 statistical components of the departments. For example, it is 
 
 certainly true that the Bureau of Labor Statistics was initially 
 established to collect statistics about and for the labor movement. 
 The Consumer Price Index, to some extent, was tailored to urban 
 labor and blue-collar workers. Recently, when there was a 
 proposal to expand coverage of the CPI to include more people, 
 that is, elderly, poor, and white-collar workers, the labor 
 movement raised a great outcry. They said that the BLS was 
 destroying a vital tool, which is used in labor negotiations. The 
 inclusion of all those other people made the CPI less useful for 
 labor. 
 
 My point is that the history of this matter illustrates a 
 struggle in statistical agencies between planning for the future and 
 meeting immediate and continuing policy needs. Over time 
 statistical programs evolve into programs with new orientations. 
 Data on program administration evolve into data on program 
 evaluation. As administrative data become available covering 
 several years, a time series develops and the data evolve into 
 general-purpose statistics. 
 
 For example, in administering the Federal prison system 
 information is collected on the number of prison cells available 
 for occupancy throughout the country; in time, those data 
 become a social indicator. Thus, general-purpose needs are 
 imposed on various special purpose activities. 
 
 A goal of statistical policy is to understand this layering of 
 needs, and to see to it that the Federal and nationwide statistical 
 programs meet as broad a cross section of these needs as possible. 
 That is not to minimize agency interests, but to encompass 
 agency interests with some other concerns. That's the nature of 
 our planning and coordination responsibility. 
 
 Student: The system that you are working to establish would 
 formulate an ultimate objective related to the value of individual 
 programs. Would this be a means for OMB to determine the 
 allocation of funds to the various departments and agencies? 
 Basically we are talking about money, aren't we, when we are 
 talking about OMB? 
 
 Dr. Duncan: Yes and no. There is a management activity as well 
 as a budget activity. 
 
 Student: I am trying to understand your goals in the context of 
 the political role of the executive branch. 
 
 Dr. Duncan: There is no question about the fact that, when we 
 get to box 4, Systemwide Objectives, we are speaking of the 
 Federal statistical system, including all the departments and the 
 executive branch of Government. What I try to do in terms of 
 defining systemwide objectives is to keep in the forefront the 
 objectives of serving the Congress and the various user groups 
 both inside and outside Government. That's very idealistic but I 
 believe it is consistent with the history of our office, and is 
 essential to meet the needs of the late 1970's and the 1980's. 
 
 There are checks and balances in this kind of operation. In 
 1972, within the Office of Management and Budget, there was a 
 desire to cut down on the Government's total budget. We went 
 through a series of items and asked each department to find 
 places where they could cut their budget. It was suggested, for 
 example, that the census of agriculture be delayed and conducted 
 at the same time as the economic censuses. It sounded like a 
 pretty reasonable idea and it would have saved money. The 
 Census Bureau went forward with a budget that did not include 
 requests for dollars to conduct the census of agriculture. When 
 
 146 
 
PLANNING THE FEDERAL STATISTICAL SYSTEM-A HISTORICAL PERSPECTIVE 
 
 this decision reached the Congress of the United States, the 
 executive branch was reminded that it was required by law to 
 conduct a census of agriculture. Therefore, the Congress man- 
 dated that the Secretary of Commerce budget the agriculture 
 survey. The check on OMB's action was, and is, that Congress 
 authorizes the dollars. 
 
 Let me make another statement about our Federal Reports 
 Act responsibility. Senator Thomas Mclntyre wrote an article for 
 Reader's Digest a couple of months ago in which he was very 
 critical of OMB's administration of the Federal Reports Act over 
 the years. 1S Because of this Congressional interest, we have 
 initiated an independent outside evaluation to accomplish three 
 objectives. 
 
 1. We want to define the total reporting burden for the 
 typical small businessman, and its source. For example, 
 there are many reports required by State and city 
 agencies. 
 
 2. We have asked the evaluator to look at our processes of 
 review and clearance. In addition to Statistical Policy 
 Division personnel, many others participate in the 
 clearance process. Budget examiners evaluate policy 
 requirements. We receive outside advice from statis- 
 ticians and the Business Advisory Council on Federal 
 Reports and so forth. Obviously, we want a critical 
 evaluation of the total process. 
 
 3. Based on these two findings we want the evaluator to 
 outline the essentials of an ideal report clearance 
 function, setting aside whatever constraints there are on 
 the system at present. He has to look at what our real 
 situation is, but he has the option to anticipate and 
 design an ideal one. 
 
 My hope is that as a result of this activity we'll come up 
 with an improvement to the report clearance process, and that the 
 resulting activity will assist the planning coordination process. 
 
 I mention reporting burden because it is one of our two 
 main legislative responsibilities. We anticipate some useful results 
 from the contractor's study, and I think it will have an impact on 
 the planning process. 
 
 Student: As regards the reporting burden, I fill out an income tax 
 form for the Internal Revenue Service and as soon as I have that 
 finished I get one from the State similar almost word for word 
 to the IRS form. I recognize that the State has no doubt done 
 quite a lot of work to standardize their procedures with the 
 Federal Government, but it doesn't make it easier for me when I 
 have to fill out two forms. 
 
 Dr. Duncan: I should like to add parenthetically that IRS forms 
 are excluded from the Federal Reports Aot and are, therefore, 
 not subject to review in my office. 
 
 The reason for the different forms tends to be legislative. In 
 the case of income taxes each State has its own legislation. We 
 hope to see State statistical offices established with a responsi- 
 bility for supervising the reporting burden similar to ours. Some 
 States already have these. Once there is a State office with that 
 responsibility, then the idea of using the same form for Federal 
 and State governments can be introduced. Still the option, under 
 
 15 Reader's Digest, April 1974, pp. 156-159. 
 
 our Constitution, is vested in the States. The State has to make 
 the final decision. Tax forms are beautiful examples of where 
 coordination could be highly effective. 
 
 Student: I am impressed with the amount of statistical informa- 
 tion that the Census Bureau accumulates. Do you think you will 
 ever get better dissemination of that information to the public? 
 In other words, will you ever take it out of the hands of the 
 politicians who only quote you the particular information 
 expedient for their reelection? 
 
 Dr. Duncan: That's a very interesting comment. The researcher 
 frequently has difficulty finding out what information is buried 
 here and there. However, I don't really think that anybody is 
 consciously trying to bury it. We have an incredibly complex 
 system which generates a lot of different data. If you know what 
 questions to ask, you can usually get the information. 
 
 We're making several attempts to improve the situation, but 
 we have a very long way to go. One is a series of publications that 
 identify what is available in the way of statistics and reports. The 
 Census Bureau has a long list of such publications. The National 
 Technical Information Service, another branch of the Department 
 of Commerce, has recently issued a catalog dealing with available 
 computer tapes. We are hoping to expand the publication of 
 catalogs and directories concerning statistical information. 
 
 Mr. Barabba: I don't think there is any sort of effort on the part 
 of any politician, at least not that I'm aware of, to withhold 
 information. As a matter of fact, we don't deal directly with the 
 politicians, we release information to the politicians on the basis 
 of an OMB circular at the very same time we announce its 
 availability to the public. 
 
 It's not a problem of anyone wanting to withhold informa- 
 tion from the public. The problem is making it available to meet 
 people's varied needs, and that becomes a very, very difficult 
 task. Other agencies within or outside the Federal Government 
 are considering this problem. One of these groups is the American 
 Library Association. 
 
 We are now working on plans to involve ourselves in the 
 education of librarians. We want them to learn how Census 
 information can assist them in doing their job, so that by the time 
 a person becomes accredited as a librarian or has completed a 
 refresher course, knowledge of the use of Census data will have 
 been a part of the curriculum. 
 
 Also, the land grant colleges have extension services funded 
 by the Department of Agriculture to improve farming and the 
 rural areas through new methodologies. As farming has become 
 more and more of a business these extension programs have 
 moved toward the problems of rural America from a planning and 
 administration point of view. The extension program of 
 Oklahoma State University has come to us with a very interesting 
 proposal. It essentially says that they have an extensive communi- 
 cation network throughout the State of Oklahoma, with exten- 
 sion agents in every county, a television network reaching class- 
 rooms and facilities set up throughout each county. What they 
 would like to do is to act as a Census agent for the distribution 
 of information, but not for the collection. They want to build a 
 program where they will actually go in and develop the best ways 
 to teach people how to use Census information. If that sort of 
 program were to happen on a national or regional basis we would 
 approach a maximum of the potential distribution. 
 
 147 
 
PLANNING THE FEDERAL STATISTICAL SYSTEM-A HISTORICAL PERSPECTIVE 
 
 In a sense, the Bureau is looking for outlets for all of its 
 information. We have finally reached the point where we are 
 fairly comfortable with the data we are produciag. But, we still 
 have a couple of problems in making data available in a useable 
 fashion, and making them sufficiently accurate when they have 
 been disaggregated for small areas. Also, we have to bring together 
 the data and the people who are interested in them. 
 
 Dr. Duncan: To avoid political manipulation we have the 
 directive which Mr. Barabba referred to; it is Circular A-91 , which 
 requires statistical agencies make advance notification as to when 
 a report will be available. 
 
 Let me give you an example of why we would want to do 
 that. Let's say that it is November in an election year. A 
 candidate might want to delay the unemployment statistics 
 because they are bad or to speed them up because they are good. 
 Well, thanks to A-91 that's not possible. The publication date for 
 the unemployment rate is announced 6 months in advance when 
 no one is able to predict what the future unemployment figures 
 are likely to be. 
 
 Student: I really didn't mean manipulation in that sense. I meant 
 politicians who in their campaigns bring out only selected 
 statistics, and the constituents don't have the knowledge to 
 differentiate. 
 
 Mr. Barabba: This is why we have more than one person running 
 for the same office. The critical element as regards information is 
 whether the opposition, potential opposition, or anyone inter- 
 ested in a particular angle or aspect of an issue, has access to the 
 information. Of course, this concept includes the provision that 
 no single person has an advantage in timing as to when they can 
 get the information. This is also extremely important. Dr. 
 Duncan's operation makes sure that everybody who controls the 
 information abides by rules governing the release of information. 
 If there is a major technical problem involved in delivering on 
 time, we have to give his office sufficient notice and show good 
 reasons for the delay. 
 
 Student: You have alluded to standarizing Federal procedures 
 and trying to integrate statistical procedures into a system that 
 reduces overlap. Do you receive any feedback from the State or 
 local governments concerning these problems? Also, what sort of 
 statistical methods do you have or hope to develop to improve 
 efficiency in making the data available in a more timely or 
 relevant fashion? 
 
 Dr. Duncan: Take something like the State Employment Service. 
 It collects wage data for administering State laws, but summarizes 
 that data for Federal authorities according to a large manual of 
 instructions, prepared by the Bureau of Labor Statistics. There is 
 a tremendous amount of standardization in programs of that 
 type. 
 
 Now there are other types of standardization. For example, 
 we have a set of criteria for defining a standard metropolitan 
 statistical area, including the level of commuting between 
 counties, the density of the areas, and so forth. We frequently 
 have interaction with States where they, for the purposes of 
 sub-State regional planning, would like to use different standards. 
 
 Then there are detailed technical assistance programs. The 
 Census Bureau has developed some methodology for making 
 interim population forecasts between censuses. The Bureau 
 
 148 
 
 actually certifies the procedures of certain States and lets them 
 carry the responsibility for making the interim forecasts. I feel 
 that we ought to do more of that sort of thing. 
 
 Mr. Barabba: The closest we have come to this is with the Center 
 for National Health Statistics as it deals with vital statistics, where 
 the center sets its own standards and then evaluates the numbers 
 provided by counties and States. 
 
 As to methodology, the Bureau is finding a much greater 
 desire by government agencies and the private sector to undertake 
 sample surveys. Also, there's a lot of pressure on the Bureau to 
 make more of its methodology known. Interestingly enough, I 
 think we already make a significant effort in this area. Informa- 
 tion on our methodologies is available in published monograph 
 form. Yet, we are today coming to believe that even that is not 
 enough. However, making methodologies available is not our 
 specific mission and, therefore, we run into a problem with 
 funding from the Congress. 
 
 Student: Is one of the central problems of regional and local 
 governments the lack of knowledge and expertise of what 
 resources are available? 
 
 Mr. Barabba: It's a massive problem! The central stumbling block 
 is that some people are exceptionally sophisticated in the use of 
 this resource, while others are very poor. Were we not faced with 
 these different levels of understanding our task would be much 
 easier. 
 
 If all counties kept their assessment records like Orange 
 County, Calif., we perhaps wouldn't need the Internal Revenue 
 Service records to conduct a census of agriculture. We could go 
 right to the assessor's file for the sample. As a matter of fact, 
 Representative Andrew Hinshaw, who used to be the assessor of 
 Orange County and now sits on the Census Oversight Committee, 
 has a dream of taking a census entirely from administrative 
 records. He believes that if you leave it up to the assessors of 
 America you can replace the Census Bureau. Whenever we do a 
 study he asks us why we aren't using administrative records. But, 
 on the other hand, another oversight committee asks us why we 
 are using administrative records at all. 
 
 At any rate, there's a wide variety of standards and talent in 
 the States. We, at the Bureau, can see the States increasing in 
 their desire to do the job and right now the States have an 
 economic surplus. It is the local areas that have financial 
 problems, and this is one of the things that will play a key role in 
 the subject of GBF files. These files will inevitably be a unifying 
 device which will allow us to combine a lot of diffuse data. 
 
 Dr. Duncan: Two additional background comments. The 
 American Statistical Association has through its local chapters 
 gone into individual States and evaluated the job descriptions and 
 the civil service requirements for State statisticians. A summary 
 of these is available and being reviewed. 
 
 Second, as a result of our regional council consultations, 
 which we are holding this month, we will probably be moving in 
 the direction of having a national conference at which time we 
 will determine which States have an interest in statistics. 
 Hopefully, we will be able to move in a direction of getting some 
 evaluation as to leading or lagging States and what can be done to 
 improve the situation. Obviously, these problems exist in depart- 
 ments within States and so we can have a situation where a State 
 is advanced in some areas and extremely poor in others. What we 
 
PLANNING THE FEDERAL STATISTICAL SYSTEM-A HISTORICAL PERSPECTIVE 
 
 are talking about at the moment is simply identifying those States 
 that have a broad interest in developing a strong coordinated 
 statistical program. 
 
 Student: Shortly our new President will be holding a summit 
 meeting on inflation and presumably on the Federal budget. 
 Could you tell us what the role of your agency will be as regards 
 that meeting? Or what objectives you will try to convey to them, 
 as an administrator? 
 
 Dr. Duncan: I have already been volunteering! I am thoroughly 
 convinced that inflation has had a very serious deteriorating 
 impact on our statistical system. As I mentioned earlier, I believe 
 that these problems were becoming evident several years ago, and 
 they are absolutely clear today. One of the things that we will be 
 doing is working through the Council of Economic Advisors and 
 through the Economic Policy Council to define priorities for 
 improving statistics about the economy. 
 
 On a broader level, the Office of Management and Budget 
 will be involved in the summit conference. Beyond doubt, a 
 primary topic of discussion is going to be the Federal budget and 
 the principles associated with keeping a balanced budget in fiscal 
 1976— an already stated goal. No doubt there will be some 
 attention paid as to what can be done in fiscal 1975, although in 
 that case things are pretty well set in place already. However, all 
 of the appropriation bills have not been signed and this allows a 
 degree of freedom. 
 
 You may be familiar with the fact that the Budget Reform 
 Act of 1974 eliminated the President's ability to impound funds. 
 For example, if the executive branch thought inflation was a 
 critical problem, funds for approved programs could be with- 
 held. To pump up the economy, funds previously impounded 
 could be released. That approach is not open to the President 
 today, and my guess would be that at that economic summit 
 there will be a rather elaborate presentation by the OMB on both 
 the nature of the budget and what it means to the decisionmaking 
 process, as well as an explanation of the current budget and what 
 actions are available. 
 
 Under current legislation, $198.4 billion of a $305 billion 
 budget is preprogrammed into areas like the Social Security, as 
 defined by law. That leaves roughly $100 billion to work with on 
 a discretionary basis, of which defense and Federal employment 
 represent a substantial portion. 
 
 The only way the Statistical Policy Division can have an 
 impact on the budget is to have the budget examiners working 
 with us. If the examiners feel our office is the mouthpiece of the 
 Census Bureau, and we are simply saying "more statistics is good" 
 without an objective perspective, they would stop listening to our 
 advice. 
 
 Student: Is the "management by objective" system (MBO) more 
 efficient than the Program Planning and Budget System? 
 
 Dr. Duncan: Yes, only because of the way the MBO process has 
 been defined to date. There is nothing inherent in the process 
 that should make it any easier. 
 
 Mr. Barabba: For example, I remember that we had a very 
 difficult time trying to redirect our budget effort toward users. I 
 assessed the main objectives of the MBO program and found an 
 objective called "The New Federalism." Then, I asked the ques- 
 tion, "Aside from giving this money back to the cities, do we 
 need information as to the most efficient way to allocate the 
 dollar?" The answer was "yes." I think it's a very efficient way to 
 tackle problems. 
 
 Student: Is there a rule as to when a questionnaire comes in to 
 your office that you're bound to give them an answer within a 
 particular period of time? 
 
 Dr. Duncan: As far as our procedures go, there is no deadline. 
 When someone comes to us for approval of a form we don't have 
 to respond within any prescribed period of time. We try to 
 handle the vast majority of requests within 2 weeks of their 
 arrival. On the other hand I just received a list from one agency 
 listing a dozen of their reports that we have had in our hands for 
 an extended period. 
 
 Many times when we ask an agency to state why it is 
 necessary to ask a certain question there is a long lag between our 
 question and their reply. The difficult thing for most people 
 viewing our system to understand is why we approve 95 percent 
 of the requests. With that high an approval rate it doesn't 
 appear that we are really doing our job. However, what that 
 approval rate doesn't reveal is that many requests are withdrawn. 
 Our system of accounting doesn't reflect the withdrawals. 
 
 Mr. Barabba: A lot of times a proposed questionnaire will come 
 across my desk and I will know that its acceptance or rejection 
 will be very close. So, my Forms Clearance Officer calls Dr. 
 Duncan's Forms Clearance Officer and many times that form will 
 never even go downtown. 
 
 Dr. Duncan: Then, too, we ultimately approve a lot of items that 
 look very, very, different from their original form. We may end 
 up approving a small agency request for a sample survey which 
 was originally sent to us as a census. A lot of the small agencies 
 get the idea that they should take a complete census of their 
 particular area of interest. 
 
 We do not try to measure the proposed reporting burden 
 coming into our office or the final burden that goes out to the 
 public, but I am certain that it is considerably reduced by our 
 review process. 
 
 Mr. Barabba: If we sent OMB a questionnaire that we felt was 
 extremely important and they began to delete questions that we 
 felt would alter the nature of the survey or make the survey 
 meaningless we would have to go through negotiations. 
 
 149 
 
PLANNING THE FEDERAL STATISTICAL SYSTEM-A HISTORICAL PERSPECTIVE 
 
 APPENDIX 
 STATISTICAL POLICY DIVISION INTERAGENCY COMMITTEES 
 
 1. Interagency Committee on Balance of Payments Statistics 
 
 2. Interagency Advisory Committee on Defense Indicators 
 
 3. Interagency Committee on Foreign Trade Statistics 
 
 4. Interagency Committee on Foreign Trade Commodity 
 Classification 
 
 5. Committee on Income Distribution Statistics 
 
 6. Technical Committee on Industrial Classification 
 
 7. Federal Committee on International Statistics 
 
 8. Interagency Committee on Output and Productivity 
 
 (formerly: Interagency Committee on Measurement of 
 Real Output) 
 
 9. Occupational Classification Committee 
 
 10. Interagency Research Committee on Seasonal Adjustment 
 
 1 1 . Federal Committee on Standard Metropolitan Statistical 
 Areas 
 
 12. Statistical Reporter Editorial Committee 
 
 13. Policy Committee on Current Population Survey 
 
 14. Technical Committee on Labor Force Statistics 
 
 15. Interagency Committee on Statistical Program and Policy 
 
 16. Interagency Committee on Financial Statistics 
 
 17. Committee on Environmental Statistics 
 
 18. Committee on Price Statistics 
 
 19. Task Force on the Quarterly Financial Report 
 
 20. Technical Committee on Poverty Statistics 
 
 21. Interagency Committee on Social Indicators 
 
 22. Interagency Committee on GNP Statistics 
 
 23. Interagency Committee on Transportation Statistics 
 
 24. Task Force on Standard Statistical Establishment List 
 
 25. Committee on Statistical Methodology 
 
 150 
 
COMMERCIAL USE OF CENSUS DATA 
 
 Max Eveleth, Jr., President, Urban Data Processing 
 
 "Another interesting thing is the mathematical calculation of the actual available 
 living space within a census tract. You look at a tract and determine that there is not 
 much room for growth. Therefore, the decision to open a restaurant surrounded by many 
 tracts with a small growth potential is a negative one. On the other hand, we find a series 
 of tracts which the model indicates over the last 3 years are growing. There's quite a bit 
 of room in terms of livable space, and density is low. Wham! A major developer puts in a 
 $28 million shopping center. These indexes and suggestions are very important." 
 
 Max Eveleth, Jr. 
 
 Mr. Eveleth: Good afternoon, I am Max Eveleth, president of 
 Urban Data Processing of Boston. I'm from the "other world," to 
 a certain extent, that is, "the capitalists and the exploiters." 
 
 Basically, during the last 2 or 2 1/2 years Urban Data 
 Processing has dedicated itself, in very simple terms, to trying to 
 present to industry a manageable way of effectively utilizing the 
 data and results of the 1970 census. There are a number of people 
 on the staff of Urban Data Processing that came out of the 
 Census Bureau's census test program conducted in New Haven, 
 Conn. These people are very knowledgeable about census data 
 and the systems that the Census Bureau has developed. 
 
 Primarily, we look upon ourselves as being in the market 
 segmentation business. For the first 3 years of Urban Data the 
 boys were developing software systems, packages, and specialized 
 subsets of the 1970 census data, with the express purpose of 
 utilizing these in a marketing environment for private industry. 
 Then, a couple of years ago, we had it all ready, and encountered 
 the first problem. We asked industry, "Gee, are you using census 
 data?" The guys said, "No, why would I want to do that?" We 
 were forced to realize that before we could get a fellow excited 
 about our fantastic resource program we really had to "turn him 
 on" to the tremendous amount of data available that could be a 
 very useful marketing information tool. 
 
 To explain the distinction between data and information, 
 let me try to draw an analogy. If you got to the floor of the New 
 York Stock Exchange at the end of a day of trading, the floor 
 would be covered with buy /sell papers. Those are data. They tell 
 you what stocks were bought and sold and how many of each. 
 The tables that come out in the Wall Street Journal the next 
 morning which sum this up and indicate price changes and shares 
 traded are information. You can use that information very 
 quickly to decide whether to maintain your holdings. 
 
 In the same way the Census Bureau has gathered an 
 enormous amount of data. The challenge is to effectively and 
 economically convert that data into information that you can use 
 as a management input for marketing decisions. 
 
 The census questionnaire is designed to learn some basic 
 information about the American people: who they are, where 
 they are, where they came from, what kind of house they live in, 
 and what facilities are available. 
 
 Census data cover a multitude of information, from the 
 number of bathrooms in the home and whether the television 
 receives a UHF signal, to how people get to work. 
 
 In an attempt to economize, the Census Bureau employed a 
 mail-out/mail-back method to conduct the 1970 census in the 
 
 145 largest standard metropolitan statistical areas (SMSA's). 
 Computers also contributed to efficiency and economy during 
 the survey. 
 
 Census data are released on several levels, ranging from a 
 single city block in some of the larger cities, and groups of city 
 blocks or their counterparts outside the metro areas called 
 enumeration districts, to SMSA census tracts, which contain 
 approximately 4,000 persons, or municipal territories outside the 
 SMSA. The information is released piecemeal in two major 
 segments. The data collected from the survey of the entire 
 population, which numbered 205 million in 1970, are dissemi- 
 nated in the three initial stages. Additional data from two sample 
 surveys, collected within the decennial census, are the last three 
 sets of information released. This information is gleaned from 
 more than 70 million questionnaires. 
 
 We, at Urban Data Processing, are attempting to demon- 
 strate to management groups the value of census-based data. We 
 show them how data have been aggregated and can be available 
 even at the block level. We're trying to give them an idea of 
 segmenting the country and how the census units were designed. 
 
 A lot of private industry people actually believe, for the 
 first time, that the government has taken into consideration the 
 interest of business and created this incredible marketing system. 
 We attempt to point out the needs of the people, and that the 
 reasons why the census was established happen to coincide with 
 those of business in terms of segmenting the markets. 
 
 One management practice that could be greatly enhanced 
 through the use of census data is direct mail advertising. Some 
 advertisers use a ZIP code to define a market area. However, such 
 use of ZIP codes, designed by the Post Office to facilitate mail 
 delivery, is a wasteful technique. Both the codes themselves 
 change, because territories of individual post offices change, and 
 the neighborhoods change, long after the families which originally 
 made the area a good market are gone. Meanwhile, the advertiser 
 may waste a small fortune advertising via ZIP codes. 
 
 However, census tracts, for which there are definite data 
 about the residents, are not defined by ZIP codes. AZIP-code 
 area in Marina Del Rey, Calif., with a mean income of $10,000 is 
 not a place where we're going to sell Porsche automobiles through 
 the mail. Don't laugh. I have customers that sell $22,000 
 automobiles through the mail with great success. 
 
 Now, let's consider the census tracts inside the Marina Del 
 Rey ZIP-code area. We find that the median income varies from 
 under $6,000 to over $20,000. If I attempt to sell Porsches 
 through the mail, I want to mail to those several contiguous 
 
 151 
 
COMMERCIAL USE OF CENSUS DATA 
 
 
 tracts within the community with the highest income level. 
 Remember, if you sell that kind of product through the mail, you 
 are probably spending $1.00 per brochure for postage, paper, 
 printing, stuffing and handling. Therefore, if you can eliminate 
 20,000 unnecessary pieces, you save $20,000, and the message is 
 going to the people you really want to contact. 
 
 The Census Bureau has released some data on the ZIP-code 
 level, but restricted this amount because of the ZIP-code 
 problems I mentioned before. 
 
 The best thing you can say about marketing via ZIP codes is 
 it's better than nothing. Most of the people we talk to quickly see 
 the benefits of working at the census tract level. 
 
 The thing that always comes up is the assumption that 
 1970 census data have no relevance in 1974. At the risk of bodily 
 injury and embarrassment to ourselves, we stage a mini-sociology 
 course which suggests people do things because of their own or 
 other people's social expectations— roles they want to play, status 
 they are seeking. Consumer characteristics can definitely be 
 correlated to particular kinds of demographics, according to the 
 Chicago School of Sociologists. 
 
 The validity of this conclusion has been studied. Basically, 
 when someone moves out of a neighborhood, he doesn't tear 
 down the house, he sells it. The person who buys a $150,000 
 home overlooking some country club in west Los Angeles with a 
 swimming pool and seven bedrooms probably didn't just get out 
 of the Army as a PFC, nor would it be a couple just retired at the 
 age of 74. It is probably a person similar to the one who just 
 moved. 
 
 One effective approach is for me to ask, "Have you ever 
 moved into a neighborhood or bought a house without knowing 
 or asking one thing about the area?" Did you ever just look at the 
 house and say, "That's fine. I don't care about the neighbors, or 
 their jobs, or their kids, or the school system. I want that house." 
 Let me assure you, that doesn't happen. Basically, when people 
 move in and out of neighborhoods, they're seeking their own 
 level. They are sustaining the kind of homogeneity we find in 
 census tracts. 
 
 Also, studies indicate there is a consistency among census 
 tracts over time. If I gave you a list of the 100 census tracts in 
 Los Angeles with the highest income level in 1970, and matched 
 it to a similar list in 1974, the percentage of tracts appearing on 
 both lists would be enormously high. Therefore, changes within 
 census tracts are important to us only insofar as they alter the 
 relationship of one tract to another. 
 
 Several companies are engaged in population projections. A 
 planning group and a list compilation company have coordinated 
 their efforts, and by April 1 of each year, they can estimate each 
 census tract's population and housing counts, based on the 
 number of automobiles and telephones in that area. 
 
 For example, if there is a census tract in Atlanta, Ga., 
 which in 1970 had 14 people who lived in 3 households, and 
 suddenly in 1973 there are 1,200 telephones installed in the tract 
 and 900 automobiles, something has happened! (We don't obtain 
 this updated information from the Census Bureau.) But, the 
 whole idea is that if there's that kind of activity we should take 
 note. 
 
 Another interesting thing being done is the mathematical 
 calculation of the actual available living space within a census 
 tract. You can look at a tract and determine that there is not 
 much room for growth. Therefore, the decision to open a 
 restaurant surrounded by many tracts with a small growth 
 potential is a negative one. On the other hand, we find a series of 
 
 152 
 
 tracts which the model indicates, over the last 3 years, are 
 growing. There's quite a bit of room in terms of livable space, and 
 the density is low. Wham! A major developer puts in a $28 
 million shopping center. These indexes and suggestions are very 
 important. 
 
 What I'd like to do is talk about the two systems we 
 developed to facilitate the use of this tremendous resource, the 
 1970 census. The first is called Area Profile Report System, 
 which identifies large portions of contiguous land and the 
 population located therein. 
 
 For instance, a branch bank is being planned in Harvard 
 Square. In the past the market was defined by plotting a map and 
 manually compiling the data of the census tracts therein— time 
 consuming and certainly less than exciting. Moreover, not all the 
 necessary data were printed. 
 
 We define the market area using half-mile radii. (Other 
 geometrically defined areas can also be used for smaller or larger 
 markets, such as police, fire, or health districts.) Using the 
 half-mile measures we can base our study on block groups. 
 (Although some data are published only at the census tract level, 
 it can be broken down mathematically to estimates for the 
 corresponding block groups.) 1 
 
 The Census Bureau has solved the problem of going through 
 all the information to find the pieces concerning those areas of 
 interest. This they have done by attaching the latitude and 
 longitude to the computerized form. 
 
 Mr. Barabba: Could you discuss the calculation of block group 
 data from census tract data? 
 
 Mr. Eveleth: Basically, at the block group level we know how 
 many houses, apartments, and people there are by block groups, 
 and the average housing values. If we take the average family 
 income, which is available at the tract level, it is our belief that 
 there is a relationship between home values and family income. 
 Based on population counts, we set up distribution factors. This 
 system is very general and doesn't need to be precise. 
 
 Mr. Barabba: Let me make a point here. We recently received a 
 letter from Senator Charles Percy inquiring about this type of 
 work. Some other firms take an automobile registration list, or 
 perhaps other listings of individuals available to private concerns, 
 and find out what enumeration district or census tract a person 
 lives in. Then they take the characteristics of that tract and do a 
 small survey, relating their survey results with the tract data. 
 They imply to their clients that they can provide individual 
 information on a large number of census characteristics about 
 each name on their lists. Sometimes the implication is that they 
 actually tap the individual census records. This implication is, of 
 course, erroneous. Although the mathematical calculation of 
 certain characteristics only statistically approaches the real 
 individuals in that tract, when these firms sell their lists, it's very 
 easy to say, "We have an auto registration list, and we've tied into 
 the Census Bureau data." When Senator Percy sees that, he is 
 legitimately concerned, because we claim that we print only 
 
 'The process is called "Fourth Count"— approximately 13,000 cells of 
 population and housing data for each of about 35,000 tracts and 35,000 
 census county subdivisions; also, about 30,000 cells of data for each 
 county. Source: 1970 Census of Population and Housing, "Publications 
 and Computer Tape Program." 
 
COMMERCIAL USE OF CENSUS DATA 
 
 aggregate statistics. That, in point of fact, is true. So he concludes 
 that someone has learned to beat the system, and we find 
 ourselves spending a lot of time explaining that these people are 
 overselling their product. Granted, they have done a good job of 
 using available data, and it is true that a specified individual is 
 more likely to have these characteristics simply because he lives 
 within a particular tract. That's good marketing. But, unfortu- 
 nately, they oversell and leave the wrong impression. 
 
 Mr. Eveleth: When you talk about auto registration lists, I'd like 
 to make a couple of comments. I think seven or eight States have 
 now enacted legislation that the automobile registration list is 
 proprietary and can't be released. You can check me on that. I 
 think there are another 10 or 12 where legislation is currently in 
 process. The interesting thing about an auto registration list is 
 that if you go to a place like New York City, only 16.5 percent of 
 the population own an automobile. There's another thing. 
 Because I own a hardware store, it's much more appropriate for 
 me, for tax purposes, to have the hardware store own my truck. I 
 don't own the truck, XYZ Hardware Store does. And that fellow 
 shows up as having no car. 
 
 The next thing you run across is the assigned values of these 
 automobiles as a characteristic of the owner. It's called "reverse 
 snobbery." This describes a fellow who makes $50,000 or 
 $60,000 a year, but refuses to spend $7,000 every 2 years to buy 
 a new car. He just drives his until it drops. This suggests the 
 advertiser should not approach him through the mail. Basically, 
 there are inherent weaknesses in their program. 
 
 Once we have created our file and added to each block 
 group its latitude and longitude, all we have to do is calculate the 
 potential bank site. The method, again using latitude and 
 longitude, of defining the distances between the bank and the 
 market members is simple geometry. 
 
 Let's return to Harvard Square. Having calculated the bank 
 site, we look in half-mile intervals at our market area and its 
 characteristics. All the appropriate data appear individually for 
 each interval, and as part of the aggregate. 
 
 These data tell us the number of men and women in the 
 area, by age, race, and marital status, as well as median age. Data 
 about housing units include the number of rental and owner- 
 occupied units, people per household, and average market values. 
 There is information by occupation and education. Average and 
 median income for families and unrelated individuals and data on 
 the civilian work force are also presented. We show how people 
 get to work and the number of cars available. 
 
 We even show how the home and water are heated. I don't 
 know who originally included that as part of our report, but 
 during a presentation in Tulsa, Okla., to a group of advertising 
 executives, one fellow got very excited. One of his accounts was a 
 22-State distributor of bottled gas, and he never had any idea of 
 the percentage of homes around the distributorships which used 
 bottled gas. 
 
 Households, by the year built and by the year the present 
 occupants moved in, are also part of the summary of findings. If I 
 open a department store and 70 percent of the people within a 
 5-mile radius at the time of the census had lived there for over 10 
 years, the chances are the neighborhood is highly stable. And 
 once I convince them to carry my credit card, the odds are 7 to 
 10 they will stay in the area, and I won't have the expense of 
 selling the idea again. 
 
 On the other hand, in the Harvard Square area you see a 
 tremendously high turnover. At the time of the census, 61 
 
 percent of the people had been there less than 2 years, which is 
 exactly what you'd expect in the middle of a college campus. At 
 that location, one would not be overly anxious to open a bank, or 
 a department store that uses credit. But it might be great for a 
 pizza parlor or hamburger stand. 
 
 Thus, in our summary of the census data and the most 
 recent population estimate, the customer gets a very comprehen- 
 sive display of the people within the distance specified. This 
 system is used by fast food chains, branch banking systems, 
 retailers, real estate developers, and other commercial interests. 
 
 I don't know anything about any one of the individuals 
 within the area defined, but it's certainly going to be less 
 expensive for me to promote my business through the mail than 
 it would be to run a television campaign all over Boston just to 
 get a hamburger stand going. 
 
 Student: Do you have any type of educational program which 
 explains the uses of the data? 
 
 Mr. Eveleth: Let me answer that a number of ways. Number one, 
 the presentation that you're seeing has become successful in 
 serving that particular purpose. The system I have described has a 
 combination sales brochure and user's guide that describes every 
 piece of data in the report, how it was obtained, and what to look 
 for. 
 
 An illustration of the small businessman using our data 
 would be the "Big Beef King" hamburger stand buying a report 
 from us and later selling the franchise for that site. In that case 
 the report may well be used by the new franchise owner for 
 marketing the product. 
 
 Student: Then, basically, you are data providers and don't really 
 serve the function of analysis? 
 
 Mr. Eveleth: Let me use different words. As I said, there's a 
 massive difference between data and information. The report I 
 give a client relates to the specific site he's concerned with, and 
 that's information. He cannot use data to make a business 
 decision, but he can use this APR to make a business decision. 
 Fundamentally, we convert data to information. We do not 
 interpret. We don't do models. We basically try to reduce the data 
 to useable information. 
 
 The second system we refer to as the ADAPT system. I no 
 longer care about contiguous parcels of geography. Now I have a 
 whole mess of names and addresses, for instance, everybody with 
 a Saks Fifth Avenue credit card in southern California. I know 
 some cardholders pay their bills more promptly than others, some 
 buy more and charge more than others. But I really don't know 
 what the "odds" are of a certain percentage living in owner- 
 occupied dwellings, or apartments, or those that are college 
 graduates. Nor do I know what percentage have an income in a 
 certain range. If I could take those names and addresses and relate 
 them to census data, it would be a giant step forward. It's not 
 certain just because a person lives in a particular tract that is 
 99-percent homeowner-occupied that he isn't one of those who 
 lives in an apartment. But it is probable that he owns his home. 
 
 The store credit manager, by correlating income and 
 education data to known cardholders, can determine which of 
 them live in the census tract with an average income of $15,000 
 or more. Realizing the store has penetrated only 20 percent of 
 those in the high income market with credit cards— which is itself 
 a very good percentage— he wants to solicit the remaining 80 
 percent. 
 
 153 
 
COMMERCIAL USE OF CENSUS DATA 
 
 We pull the names, addresses, and phone numbers of all the 
 residents of that census tract and match them against those 
 known cardholders. These "qualified prospects" can be checked 
 against files in a national credit bureau. Then the store can 
 contact, by mail or telephone, the credit-approved prospects. 
 
 If the store management is considering a new store, the 
 planners can take a census tract map and mark the greatest 
 penetration of credit cards. Opening the store in the middle of 
 that penetration is practical, considering that if those customers 
 spent that much money before when they had to travel to the 
 store, it's probable they are going to patronize the new store and 
 possibly spend greater amounts. 
 
 Here the uniquely important point is that I am going to 
 take the files of a department store which contain a great deal of 
 personal information about the individuals and their activities 
 with that store, relate them to census data, and return them to 
 the store. But, I don't sell that information to anyone else. 
 
 This correlation is referred to as geocoding. It is the merger 
 of various files based on a geographic code, such as an address. 
 Put simply, even ZIP coding mail is a form of geocoding. The 
 census coding is more sophisticated. 
 
 It is also possible to generalize about certain characteristics 
 of persons living outside an SMSA. But it is important to be 
 careful not to compare data from different sized census units. 
 
 In order to geocode data— matching an address to some 
 information so it can be packaged to suit a particular census 
 unit— street directories are required. Originally, the systems to do 
 this were called address coding guides, but now they have become 
 dual independent map encoding (DIME) files, a component of 
 geographic base files (GBF). 
 
 From a matching viewpoint these systems are similar. With 
 an address such as 400 Adams St., Quincy, Mass., the computer 
 goes to that section of the Boston address coding guide and finds 
 the range of Adams Street where 400 is, in terms of census tract. 
 Now, the resident is a member of a group, about which we have a 
 massive amount of demographic data. However, the person is still 
 not individually identifiable. 
 
 There are certain limitations in this system, because certain 
 areas have not been geocoded. Also, old streets are changing and 
 new ones constructed, so the maintenance of the files is a 
 continuing process. 
 
 A national women's magazine serves as an illustration of why 
 the sort of thing I'm describing is important in private industry. 
 This particular magazine has two editions: a main edition and a 
 second which is allegedly mailed to those subscribers that have 
 the highest average income. If I'm selling scotch, I know from my 
 research that 98 percent of the people who drink scotch have an 
 income over $14,500. If I can just advertise in this small section it 
 doesn't cost as much, and it is reaching a high income group. 
 That's basically the reason behind the concept of multiple 
 editions. 
 
 This particular magazine created its high income edition 
 based upon a ZIP code. We took one of those ZIP codes, Melrose, 
 Mass., as an example because it has four census tracts within it. In 
 this particular case, the magazine had 900 subscribers in the area, 
 and they wanted to evaluate whether those 900 readers should 
 receive the high income edition. 
 
 Melrose is a middle-income type of community. It's a very 
 firm community in the sense that there are probably only five or 
 six lots suitable for new construction in that city. The average age 
 of the homes is perhaps 1 5 to 18 years. Let's get to know Melrose 
 a little better. 
 
 154 
 
 This is the first count, 100-percent data for the city of Melrose. 
 Notice that in addition to just giving straight answers to the 
 question, the Bureau has, in fact, created a number of tables. In 
 table 1 we can see the relationship between black or white occu- 
 pants relative to number of rooms, availability of plumbing facili- 
 ties, and so on. 
 
 This second table displays more data describing Melrose. 
 This is fourth-count data for the entire city of Melrose, where the 
 average family income is $13,900. 
 
 We also get a rough idea of education— people who attend 
 college, male and female. There are about 2,800 people with a 
 very high income in the city. We've got 10,200 homes and 33,000 
 people, right on the national average of about 3.1 persons per 
 household. This is the level at which the publisher made his 
 decision. Let's look at the individual census tracts for the same 
 city (table 3). 
 
 The first census tract (3361) has an average family income 
 of $13,700. Everything is nicely balanced. The next census tract 
 (3362) has an average income of $17,515, 25 percent higher than 
 the rest of that ZIP code— than the rest of that city. There they 
 are, gentlemen— the "fat cats" all together. Of the 1,000 
 subscribers the publisher has in this ZIP code, there are over 600 
 in this tract alone. Notice that the people in the very high income 
 groupings, those with high levels of education and those engaged 
 in the professional occupations, are concentrated here. 
 
 If, in fact, these are critical elements for a marketing 
 campaign, and those are the only people you want, you just mail 
 to that tract. All you need now are the names and addresses for 
 the people in that tract. 
 
 This is the next tract, 3363. Our income goes down to 
 $12,700. In the last tract we drop off to $12,400. 
 
 Student: How do you relate the census data to the name and 
 address of a person? 
 
 Mr. Eveleth: To keep this simple, I have keypunched telephone 
 directories and entered them into the computer, by name, 
 address, and phone number. I matched the 52 million names and 
 addresses supplied by the Valentine-Sherman Company against 
 the GBF's for the United States and assigned to each particular 
 person, if possible, the tract code. That's now in his record. When 
 I say, "I want all the names and addresses that I have in tract 
 3748," I go into the computer, and I pull that information. Now, 
 I won't get everybody who's actually there, because there are 
 unlisted phones, people move, and others live in a tract that is 
 geocoded. 
 
 Student: You have no way of knowing somebody's income at a 
 particular address? 
 
 Mr. Eveleth: That's correct. All I know is that one person lives in 
 a tract where the average income is $10,000, and another lives in 
 a tract where the average income is $20,000. Therefore, if I'm 
 going to market Playboy Club keys, I'm only contacting the 
 fellow in the highest income tract. This means we lose some 
 potential members in the $10,000 tract, but our record proves 
 that this approach gains more than it loses. 
 
 Mr. Barabba: In a sense, the Census Bureau becomes that 
 denominator, and the local data files become the numerator, 
 because you're using local administrative records to identify 
 
COMMERCIAL USE OF CENSUS DATA 
 
 Table 1 
 
 
 
 
 
 
 
 
 I 24 POPULATION 65 YEARS 
 
 
 
 - 1 
 1 
 
 
 1 ■"" 
 
 
 
 
 
 
 
 
 I ST ATE OF MMijAl.HU 
 
 iLTT i ** 
 
 MCD= 155 
 
 MELROS F 
 
 CITY 
 
 I OLO AND OVER 
 
 I BY RELATI CNSHI P 
 
 PERSONS 
 6b YEARS 
 
 I 
 I 
 
 
 
 
 
 
 
 
 
 
 I 1 TJTAL 
 
 
 1 
 
 2 TLTAL 
 
 
 
 I 
 
 AND OVER 
 
 I 
 
 
 I PJPULATIUM 
 
 33 ,180 
 
 I 
 
 MLLSINb UNITS 
 
 10,219 
 
 
 
 I 
 
 
 
 
 1 
 
 
 
 
 
 I HEAD OF FAM LY 
 
 1 , 
 
 436 
 819 
 
 i 
 
 I 
 
 
 
 
 
 
 
 
 
 
 
 I 19 NtuKQ 
 
 AND OTHER 
 
 RACES BY AGE 
 
 
 
 I OTHER FAMILY MEMBERS 
 
 
 688 
 
 I 
 
 
 I 
 
 
 
 
 I 
 
 
 
 I MALE PRIfA"Y INDIVIDUAL 
 
 
 141 
 
 1 
 
 
 I NE 
 
 U R0 NEGRO 
 
 OTHER 
 
 OT HER I 
 
 
 
 I FEMALE PRIMARY INDIVIDUAL 
 
 909 
 
 I 
 
 
 I PALE FEMALF 
 
 MALE 
 
 FEMALE I 
 
 
 
 I NUN-KELAU VE (BOAROERSJ 
 
 
 120 
 
 ! 
 
 
 I UNDER b 
 
 8 
 
 2 
 
 5 
 
 13 I 
 
 
 
 I INMATE OF I NSTI TUTI ON - 
 
 
 111 
 
 I 
 
 
 I 3-14 
 I 15-24 
 
 I 2 b- 34 
 
 5 
 5 
 
 9 
 
 3 
 
 1 1 
 3 
 
 11 
 
 14 
 
 12 
 
 27 I 
 10 I 
 16 I 
 
 
 
 I OTHER IN oROUP QUARTERS 
 
 
 12 
 
 I 
 
 
 
 
 30 COUNT OF OCCJPIED PERSONS PFR ROOM 
 
 I 
 
 I 3b- 44 
 
 4 
 
 2 
 
 14 
 
 10 I 
 
 
 
 JNITS BY PERSONS 1.00 
 
 1.01 
 
 
 1.51 
 
 I 
 
 I 4b- b4 
 
 
 
 3 
 
 6 
 
 5 I 
 
 
 
 PER ROOM, TENURE OR 
 
 TO 
 
 
 OR 
 
 i 1 
 
 I bb-64 
 
 6 
 
 '■* 
 
 s 
 
 3 I 
 
 
 
 AND RACE OF HEAD LESS 
 
 1.50 
 
 
 MOPE 
 
 I 
 
 I ob* 
 
 2 
 
 2 
 
 4 
 
 6 I 
 
 
 
 
 
 
 
 1 
 
 1 TOTAL 
 
 J 9 
 
 30 
 
 71 
 
 90 I 
 
 
 
 TOTAL CCCJPIEC UNITS 9,646 
 
 341 
 
 
 56 
 
 I 
 
 
 
 I 
 
 18 POPULATION 
 
 Y I 
 
 OWNER OCCJPIED UNITS 6,748 
 RENTER OCCJPIED UNITS 2,898 
 
 253 
 
 88 
 
 
 22 
 
 J 4 
 
 I 
 
 ! 
 
 
 
 I 
 
 AGE ANu St \ 
 
 
 
 2 TOTAL 
 
 
 I 
 
 
 
 
 
 TOTAL NEGRO OCCJPIED 18 
 
 1 
 
 
 
 
 I 
 
 H U S 1 NG 
 
 LC ,219 
 
 I 
 
 
 MALE 
 
 -EMALc I 
 
 NEGRO OWNER OCCUPIED 10 
 
 n 
 
 
 
 
 I 
 
 UNI IS 
 
 
 IJNDE« b 
 
 1, 33V 
 
 1 
 
 2 7;> I 
 
 ^72 I 
 
 NEGRO RENTER OCCJPIED 8 
 
 i 
 
 
 r 
 
 ! 
 
 
 
 I 
 
 •> 
 
 ^9-i 
 
 
 
 
 
 
 
 I 4JA COUNT OF 
 
 I 
 
 6 
 
 33d 
 
 
 328 I 
 
 46 OCCUPIED UNITS WITH 1.01 I 47 
 
 HOUSING 
 
 .NITS 
 
 ll 
 
 I CWNFR OCCUPIED 
 
 ! 
 
 7-9 
 
 1 , uUo 
 
 
 ^bo I 
 
 OR MORE PERSONS PER ROOM I 
 
 
 
 »ITH 
 
 ! 
 
 I UNI TS WI TH 
 
 I 
 
 n-i3 
 
 1, 3o9 
 
 1 
 
 , JbV. I 
 
 HUSBANO-WIFF FAMILY 368 I 
 
 4? 
 
 
 1.51 
 
 I 
 
 
 ALL 
 
 I 
 
 14 
 
 i^l 
 
 
 369 I 
 
 OTHER FAMILY-MALF HEAD 3 I 
 
 1 
 
 OR 
 
 MORE 
 
 I 
 
 I 13 PLUMBING 
 
 I 
 
 15 
 
 j40 
 
 
 33d I 
 
 FAMILY W/ FEMALE HEAD 19 I 
 
 6 
 
 PERSONS 
 
 I 
 
 I 78 FACI LI TIES 
 
 1 
 
 16 
 
 3u7 
 
 
 3 JC I 
 
 PRIMARY INDIVIDUALS 7 I 
 
 7 
 
 PER 
 
 ROOM 
 
 I 
 
 I 31 r> poouMirw 
 
 , 
 
 I 1 
 L8 
 
 293 
 Hi 
 
 
 310 I 
 ^ttb I 
 
 
 
 
 
 
 I 1,157 
 
 VALUE 
 
 1 
 
 I 
 
 
 48 POPULATION IN UNITS WITH I 49 
 
 POPULATION IN 
 
 I 
 
 I 1,687 
 
 I S 
 
 I 
 
 19 
 
 21o 
 
 
 <l7j I 
 
 1.0 1 OR MORE PERSONS I 
 
 
 UNITS 
 
 I 
 
 I 2,C 1 6 TABULATED 
 
 I 
 
 20 
 
 ld3 
 
 
 <-l9 I 
 
 PER ROOM I 
 
 WI TH 
 
 1.51 
 
 I 
 
 I bl9 
 
 
 I 
 
 2 I 
 
 181 
 
 
 ^3 V I 
 
 A UNDER 18 1, 723 I 
 
 181 
 
 OR 
 
 MOPE 
 
 1 
 
 I 90 
 
 
 1 
 
 22-24 
 
 b90 
 
 
 770 I 
 
 G 18 - 64 996 I 
 
 113 
 
 PERSONS 
 
 I 
 
 
 
 I 
 
 25-34 
 35-44 
 
 1, 191 
 
 1 
 
 ,817 I 
 ,901 I 
 
 F 65 AND OVER 68 I 
 
 35 
 
 PER 
 
 ROOM 
 
 ! 
 
 
 
 1 
 
 1 , ?49 
 
 1 
 
 
 
 
 
 
 I -»-»A COUNT OF 
 
 
 I 
 
 45-bt 
 
 1, /89 
 
 2 
 
 , j2L I 
 
 POPULATION AND JNITS BY WITH ALL 
 
 LACKING 
 
 I 
 
 I KcNTER OCCUPIED 
 
 
 I 
 
 bb-b9 
 
 d2 / 
 
 
 9<t9 I 
 
 PLUMBING FACILITIES PLUMBING 
 
 PL UMB ING 
 
 1 
 
 1 UNITS WITH 
 
 
 I 
 
 60-61 
 
 30o 
 
 
 ->01 I 
 
 
 
 
 
 I 
 
 I ALL 
 
 
 I 
 
 62-64 
 
 -,34 
 
 
 3b8 I 
 
 51 PERSONS IN OCCUPIED UNITS 32,559 
 
 
 219 
 
 I 
 
 I 18 PLUMBING 
 
 
 I 
 
 6b-74 
 
 sat 
 
 1 
 
 ,bb7 I 
 
 52 PERSONS IN UNITS *l I T H I. 01 
 
 
 
 
 I 
 
 I 32 FACILI TIES 
 
 
 I 
 
 7 5+ 
 
 b(J4 
 
 1 
 
 , 102 I 
 
 OR MORE PERSONS PER ROOM 
 
 2, 743 
 
 
 44 
 
 I 
 
 I 252 FOR WHICH 
 I 486 VALUF 
 
 
 S 
 
 TOTAL 
 
 15,430 
 
 I 1 
 
 , 7bO I 
 
 53 COUNT OF FAMILIES 
 
 8,267 
 
 
 48 
 
 F 
 
 I ^21 IS 
 
 I 728 TABULATED 
 
 I 
 
 
 1 
 
 28A ROOMS 
 
 IN 
 
 I 28B ROOMS IN ALL UNITS 41 HOUSING UNITS 
 
 I 
 
 I 
 
 I b9^ 
 
 I URBAN 
 
 DATA I 
 
 OCCUP 
 
 IED AND 
 
 I COUNT OF WITH 
 
 ALL 
 
 LACK ING 
 
 I 
 
 I 231 
 
 I PROCESSING, IVACANT YEAR 
 
 -ROUNDl TOTAL OCCUPIED ROOMS PLUMBING 
 
 PLUMBING 
 
 i 
 
 I 8 
 
 I INCORPORATED I 
 
 HOUS ING 
 
 UN ITS 
 
 I AND VACANT 
 
 
 
 
 1 
 
 I 87 
 
 I 675 MASS 
 
 . AVE I 
 
 
 
 
 I YEAR-ROUND 60,127 10 
 
 ,100 
 
 
 112 
 
 i 
 
 
 I CAMBRIDGE, I 
 
 I ROOM 
 
 
 147 
 
 I TOTAL OCCUPIED 59,288 9 
 
 ,936 
 
 
 107 
 
 1 
 
 
 ■J 
 
 
 1 MASSACHUSETTS I 
 
 2 ROOMS 
 
 
 301 
 
 I OWN" OCCUPIED 47,083 6 
 
 ,979 
 
 
 44 
 
 I 
 
 
 I 
 
 
 12139 I 
 
 3 ROOMS 
 
 
 818 
 
 I RENTER OCCUPIED 12,2 05 2 
 
 ,957 
 
 
 63 
 
 I 
 
 
 I 
 
 
 I 
 
 4 ROOMS 
 
 
 I, 119 
 
 I NFCPO OCC UPIEO 100 
 
 19 
 
 
 
 
 I 
 
 
 I COPYRIGHT 
 
 5 ROOMS 
 
 
 1, 474 
 
 I NEGRO OWNER 70 
 
 10 
 
 
 
 
 ! 
 
 
 I 1971 
 
 I 
 
 6 ROOMS 
 
 
 2, 747 
 
 I NEGRO RENTER 30 
 
 9 
 
 
 
 
 I 
 
 
 
 
 
 7 ROOMS 
 
 
 I, 780 
 
 I VACANT FOR SFNT 2 77 
 
 64 
 
 
 3 
 
 I 
 
 
 
 
 
 
 
 
 I 
 
 I 
 
 8 PLUS 
 
 
 1, 826 
 
 I VACANT FOR SALE 156 
 
 2 2 
 
 
 
 
 I 
 
 I 
 
 indexes. Now there have been developed different indexes relative 
 to marketing. 
 
 of losing one of those tapes. But I do not incorporate the client's 
 data into our system. 
 
 Student: What do you do with the tapes when you are finished 
 with the information? 
 
 Mr. Eveleth: I send them back to the client, and I live in sheer terror 
 
 Mr. Barabba: Incidentally, a number of people from industry 
 suggested certain questions be included in the census question- 
 naire. If they had a good point, the Bureau would listen. The 
 lobbyist for the national pet food industry suggested a count of 
 
 155 
 
COMMERCIAL USE OF CENSUS DATA 
 
 Table 2 
 
 ] 1970 CENSUS OF POPULATION AND HOUSING 
 
 FOURTH-COUNT SUMMARY DATA (SAMPLE DATA) 
 
 URBAN DATA PROCESSING 
 675 MASSACHUSETTS AVE 
 
 j *** FOURTH-COUNT POPULATION TALLIES 
 
 *** 
 
 CAMBRIDGE, MASS. 02139 1 
 
 I STATE OF MASSACHUSETTS MCD = 155 = 
 
 MELROSE POPULATION = TOTAL POPULATION 
 
 COPYRIGHT 1972 
 
 ] 76 FAMILY INCOME 
 
 
 
 76 INCOME OF UNRELATED 
 
 
 42 POPULATION 
 
 25 YEARS OLD & 
 
 OVER 
 
 
 
 
 
 
 INDIVIDUALS 14 
 
 YEARS 
 
 
 BY YEARS OF 
 
 SCHOOL COMPLETED & SEX 
 
 
 
 
 COUNT OF 
 
 OLD 
 
 AND OVER 
 
 COUNT OF 
 
 
 
 
 
 
 
 FAMILIES 
 
 
 
 INDIVIDUALS 
 
 
 MALE 
 
 FEMALE 
 
 I UNDER $1 
 
 , 
 
 300 
 
 
 98 
 
 UNDER $] 
 
 ,000 
 
 
 275 
 
 
 
 
 ! $ 1,000 
 
 - 
 
 $ 1,999 
 
 
 90 
 
 $ 1,000 
 
 - $ 1,999 
 
 
 324 
 
 NO SCHOOL YRS 
 
 
 
 | $ 2 ,000 
 
 - 
 
 $ 2,999 
 
 
 130 
 
 $ 2,000 
 
 - $ 2,999 
 
 
 271 
 
 COMPLETED 
 
 25 
 
 34 
 
 I $ 3,000 
 
 - 
 
 $ 3,999 
 
 
 162 
 
 $ 3,000 
 
 - $ 3,999 
 
 
 195 
 
 ELEMENTARY- 
 
 
 
 i $ 4,000 
 
 - 
 
 $ 4,999 
 
 
 227 
 
 $ 4,000 
 
 - $ 4,999 
 
 
 193 
 
 1-4 YEARS 
 
 63 
 
 99 
 
 j $ 5,000 
 
 _ 
 
 $ 5,999 
 
 
 272 
 
 $ 5,000 
 
 - $ 5,999 
 
 
 194 
 
 5-6 YEARS 
 
 197 
 
 176 
 
 i $ 6,000 
 
 - 
 
 $ 6,999 
 
 
 302 
 
 $ 6,000 
 
 - $ 6,999 
 
 
 135 
 
 7 YEARS 
 
 146 
 
 126 ! 
 
 ! S 7,000 
 
 - 
 
 $ 7,999 
 
 
 401 
 
 $ 7,000 
 
 - $ 7,999 
 
 
 148 
 
 8 YEARS 
 
 463 
 
 630 
 
 j $ 8,000 
 
 - 
 
 $ 8,999 
 
 
 527 
 
 $ 8,000 
 
 - $ 8,999 
 
 
 124 
 
 HIGH SCHOOL- 
 
 
 
 i $ 9,000 
 
 _ 
 
 $ 9,999 
 
 
 550 
 
 $ 9,000 
 
 _ $ 9 999 
 
 
 68 
 
 1-3 YEARS 
 
 1,475 
 
 1,515 
 
 ! $10,000 
 
 - 
 
 $11 ,999 
 
 1 
 
 , 151 
 
 $10,000 
 
 - $11,999 
 
 
 96 
 
 4 YEARS 
 
 2,887 
 
 5,275 
 
 ] 512,000 
 
 - 
 
 $14,999 
 
 1 
 
 ,411 
 
 $12,000 
 
 - $14,999 
 
 
 114 
 
 COLLEGE- 
 
 
 
 i $15,000 
 
 - 
 
 $24,999 
 
 2 
 
 ,210 
 
 $15,000 
 
 - $24,999 
 
 
 49 
 
 1-3 YEARS 
 
 1,272 
 
 1,457 
 
 ! $25,000 
 
 _ 
 
 $49,999 
 
 
 631 
 
 $25,000 
 
 - $49,999 
 
 
 12 
 
 4 YEARS 
 
 1,065 
 
 801 
 
 j $50,000 
 
 AND OVER 
 
 
 57 
 
 $50,000 
 
 AND OVER 
 
 
 
 
 5 YEARS 
 
 781 
 
 290 
 
 j 68 EMPLOYED POPULATION 14 YEARS AND OLDER 
 
 
 
 
 BY OCCUPATION AND SEX 
 
 MALE 
 
 FEMALE r 
 
 I PROFESSIONAL, TECHNICAL & KINDRED WORKERS 
 
 1 ,719 
 
 1 
 
 ,034 j 
 
 ! MANAGERS & ADMINISTRATORS, EXCEPT FARM 
 
 1, 178 
 
 
 180 i 
 
 j SALES WORKERS 
 
 807 
 
 
 350 J 
 
 I CLERICAL AND KINDRED WORKERS 
 
 868 
 
 2 
 
 ,218 ' 
 
 ! CRAFTSMEN, FOREMEN, AND KINDRED WORKERS 
 
 1 ,522 
 
 
 56 ! 
 
 j OPERATIVES, EXCEPT TRANSPORT 
 
 501 
 
 
 433 | 
 
 I TRANSPORT EQUIPMENT OPERATIVES 
 
 363 
 
 
 j 
 
 ! LABORERS, EXCEPT FARM 
 
 314 
 
 
 15 i 
 
 | FARMERS AND FARM MANAGERS 
 
 4 
 
 
 o ; 
 
 I FARM LABORERS AND FOREMEN 
 
 
 
 
 ] 
 
 ! SERVICE WORKERS, EXCEPT PRIVATE HOUSEHOLD 
 
 597 
 
 
 624 ! 
 
 j PRIVATE HOUSEHOLD WORKERS 
 
 8 
 
 
 44 i 
 
 I OCCUPATION NOT REPORTED 
 
 350 
 
 
 264 i 
 
 1 AGGREGATE & FAMILY INCOME 
 OF FAMILIES 
 
 AVERAGE FAMILY INCOME ($$) 
 
 3 AGGREGATE $ INCOME OF UNRELATED 
 INDIVIDUALS 14 YEARS & OVER 
 
 AVERAGE UNRELATED INDIVIDUAL INCOME 
 
 114539750 
 
 13,936 
 
 11177700 
 
 5,083 
 
 I ' 
 
 126 100% POPULATION COUNT 
 
 33, 180 
 
 127 100% HOUSING UNIT COUNT 10,219 
 
 the number of pets in a household. For the pet food industry that 
 information would have been fantastic. But, there was no way the 
 Bureau could justify such a question on the census form. 
 
 As I said earlier, it would have been ridiculous at one time 
 to ask about plumbing facilities. Today, it's a much more 
 legitimate question. Natural resources was another such question, 
 but in 1970 we dealt extensively with how people got to work. 
 Faced with diminished oil supplies, society has decided to find 
 new modes of transportation. So, you need data on the local 
 level, and we value suggestions from this level. 
 
 Regarding advertising through the mail, we, in this class, 
 have mixed emotions about computer letters. 
 
 Mr. Eveleth: I'm offended by some of the ones I get, because I 
 feel they think I'm an idiot. If they think that, because they 
 repeat my name four times in a letter, I'm stupid enough to 
 believe I got a personal letter, good luck to them. When I say a 
 computerized letter, I am speaking of the way we do it, where the 
 entire letter is printed on the computer. That's not what they call 
 
 a "filler." It really looks like a typed letter. We use a special kind 
 of print chain. The ink we use to stamp the signature, if you wet 
 your finger and rub it, will smudge. One of the problems we have 
 in judging the success of computer letters is that some people 
 have refined the science to the point that the person has no idea 
 it's a computer letter. 
 
 The contrasting reaction is getting a letter that says "Dear 
 Customer," and offering me a charge card when I've had one 
 with that same bank for 3 years. Then I wonder what kind of 
 bank I do business with? 
 
 Student: Are you making this information available to consumer 
 groups? Say, one of them wants to make a mailing to cover 
 Marina Del Rey. Has anybody used your services? 
 
 Mr. Eveleth: Yes. 
 
 Student: Has it been successful? 
 
 156 
 
COMMERCIAL USE OF CENSUS DATA 
 
 Table 3 
 
 Census Tracts 
 
 INCOME IN 1969 Of FAMILIES AND 
 UNRELATED INDIVIDUALS 
 
 *■!•«■•■ 
 
 Less thon $1,000 
 
 SI 000 lo $1,999 
 
 $2 000 to $2.999.. 
 
 S3 000 lo $3 999. 
 
 $4,000 lo $4.999 _ 
 
 $5,000 lo $5.999 
 
 $4,000 10 $6,999.. 
 
 $7,000 lo $7.999 
 
 $8 000 to $8.999. , 
 
 $9,000 lo $9.999 
 
 $10.000 10 $11.999 
 
 $12,000 10 $14.999 
 
 $15.000 10 $24.999... 
 
 $25,000 lo $49 999. _ 
 
 $50,000 or more 
 
 Medion income 
 
 Meon income 
 
 Families ond unrelated individuals 
 
 Medion income 
 
 Mean income 
 
 Unrelated individuals _ 
 
 Medion income 
 
 Meon income 
 
 SCHOOL ENROLLMENT 
 
 Enroled etrtoni, 3 te 34 yeeri old 
 
 Nursery school 
 
 Public __ 
 
 Kindergarten 
 
 Public 
 
 flementory 
 
 Public 
 
 High school 
 
 Public 
 
 College 
 
 Percenl enrolled in school by aoe 
 
 16 ond 17yeors 
 
 18 ond 19 yeors 
 
 20 ond 21 yeors 
 
 22 To 24 yeors 
 
 25 to 34 yeors 
 
 Percent 16 to 21 yeors not high school graduates 
 ond not enrolled in school 
 
 YEARS OF SCHOOL COMPLETED 
 
 Partem, iS yean oM o.. em 
 
 No school yeors completed 
 
 Elemenlory I to 4 yeors 
 
 5 to 7 yeors 
 
 8 yeors 
 
 High school I to 3 yeors 
 
 4 yeors 
 
 College I lo 3 yeors 
 
 Medion school yeors completed .. 
 
 Percenl high school groduotes 
 
 
 Mel 
 
 rose 
 
 
 Trod 
 
 1 1 U< 1 
 
 Trod 
 
 Tract 
 
 3361 
 
 3362 
 
 3363 
 
 3364 
 
 1 719 
 
 1 1(1 
 
 2 043 
 
 3 S77 
 
 22 
 
 II 
 
 42 
 
 23 
 
 16 
 
 20 
 
 19 
 
 15 
 
 19 
 
 21 
 
 3V 
 
 51 
 
 39 
 
 5* 
 
 37 
 
 SO 
 
 25 
 
 44 
 
 35 
 
 123 
 
 47 
 
 25 
 
 69 
 
 131 
 
 60 
 
 59 
 
 64 
 
 1 19 
 
 83 
 
 87 
 
 142 
 
 89 
 
 99 
 
 71 
 
 142 
 
 215 
 
 137 
 
 79 
 
 138 
 
 196 
 
 244 
 
 216 
 
 298 
 
 19 J 
 
 305 
 
 235 
 
 419 
 
 452 
 
 515 
 
 618 
 
 500 
 
 5/7 
 
 93 
 
 126 
 
 93 
 
 1 19 
 
 15 
 
 33 
 
 5 
 
 4 
 
 $12 674 
 
 $15 591 
 
 $11 973 
 
 $11 305 
 
 $13 719 
 
 $17 515 
 
 $12 769 
 
 $12 393 
 
 2 177 
 
 2 204 
 
 » 346 
 
 3 691 
 
 $11 012 
 
 $13 730 
 
 $11 056 
 
 $9 301 
 
 $11 861 
 
 $15 880 
 
 $11 606 
 
 $10 205 
 
 458 
 
 323 
 
 304 
 
 1 114 
 
 $3 774 
 
 $5 208 
 
 $2 837 
 
 $4 521 
 
 $4 887 
 
 $6 360 
 
 S3 795 
 
 $5 145 
 
 2 I4S 
 
 2 471 
 
 2 aid 
 
 2 653 
 
 29 
 
 64 
 
 44 
 
 66 
 
 14 
 
 7 
 
 
 Hi 
 
 125 
 
 i 15 
 
 166 
 
 48 
 
 110 
 
 129 
 
 166 
 
 143 
 
 1 213 
 
 1 331 
 
 1 604 
 
 1 402 
 
 1 176 
 
 1 163 
 
 1 520 
 
 07 3 
 
 567 
 
 656 
 
 827 
 
 656 
 
 540 
 
 585 
 
 782 
 
 61 
 
 211 
 
 285 
 
 173 
 
 1HO 
 
 88 8 
 
 99 9 
 
 999 
 
 98 9 
 
 84 9 
 
 54 1 
 
 61 
 
 56 6 
 
 28 
 
 71 5 
 
 21 ) I 
 
 16 2 
 
 17 2 
 
 19 8 
 
 63 
 
 5 
 
 6 1 
 
 80 
 
 54 
 
 
 10 3 
 
 23 
 
 6 1 
 
 '. 2 
 
 3 4SJ 
 
 4 299 
 
 4 55S 
 
 5 965 
 
 
 5 
 
 35 
 
 19 
 
 22 
 
 25 
 
 15 
 
 100 
 
 74 
 
 85 
 
 \'n, 
 
 290 
 
 182 
 
 202 
 
 J .18 
 
 371 
 
 471 
 
 510 
 
 918 
 
 1 091 
 
 1 958 
 
 1 645 
 
 1 969 
 
 2 590 
 
 618 
 
 842 
 
 585 
 
 684 
 
 633 
 
 985 
 
 499 
 
 82(1 
 
 126 
 
 128 
 
 124 
 
 12 4 
 
 81 1 
 
 8fi8 
 
 67 
 
 686 
 
 Census Tracts 
 
 KC UP AT ION 
 
 Ttfil •wpkyxJ. 16 .car. oW mni »tr 
 
 Professional, techmcol and k.ndred workers 
 
 Health workers 
 
 Teachers elementary and secondary schools . . . 
 Managers and admm.strolors. except form 
 
 Soloned 
 
 Sell employed m ret oil trade 
 
 Soles workers 
 
 Retail trade 
 
 Clerical and k.ndred workers 
 
 Oottsmen foremen and kindred workers 
 
 Mec homes ond repairmen 
 
 Operolives e»cept transport _. 
 
 Transport equipment operatives 
 
 Laborers, except farm 
 
 form workers 
 
 Service workers' 
 
 C Icon in g ond food service workers 
 
 Personol ond heolth service workers 
 
 Pnvote household workers. 
 
 lo-oU tmpbyti, 16 fwmn «M »n«J «*«r_. . 
 
 Professional techmcol. ond k.ndred workers 
 
 Teochers elementory ond secondary schools . . . 
 
 Managers and odmmislrotors except form 
 
 Soles workers 
 
 Cler.col ond k.ndred workers 
 
 Secretaries stenogrophers. ond typists 
 
 Operolives, including tronsport 
 
 Other blue-collar workers 
 
 Form workers 
 
 Service workers e«cept private household 
 
 Private household workers 
 
 TYPE OF INCOME IN 1969 OF FAMILIES 
 
 Allwlli 
 
 With wage or salary income 
 
 Meon woge or solary income 
 
 With nonform self-employment income 
 
 Meon nonform self employment income 
 
 With farm self employment income 
 
 Meon form self -employment income 
 
 With Socwl Security income 
 
 Meon Socwl Security income 
 
 With public ossistonce or public welfare income .. 
 
 Meon public ossistonce or public welfare income 
 With other income 
 
 Meon other income __ _ 
 
 Melrose 
 
 Trod 
 
 Iroit 
 
 
 rad 
 
 Iroct 
 
 3361 
 
 3362 
 
 
 112,1 
 
 3364 
 
 3 933 
 
 3 907 
 
 3 
 
 160 
 
 4 307 
 
 650 
 
 741 
 
 
 496 
 
 9 78 
 
 82 
 
 137 
 
 
 76 
 
 7 1 
 
 151 
 
 126 
 
 
 80 
 
 '.' 
 
 372 
 
 452 
 
 
 271 
 
 12 
 
 348 
 
 437 
 
 
 251 
 
 293 
 
 8 
 
 10 
 
 
 14 
 
 19 
 
 273 
 
 298 
 
 
 286 
 
 D > 
 
 105 
 
 99 
 
 
 155 
 
 155 
 
 710 
 
 641 
 
 
 733 
 
 1 083 
 
 292 
 
 290 
 
 
 456 
 
 622 
 
 64 
 
 65 
 
 
 124 
 
 161 
 
 82 
 
 85 
 
 
 126 
 
 58 
 
 215 
 
 203 
 
 
 286 
 
 137 
 
 83 
 
 41 
 
 
 116 
 
 ll 
 
 73 
 
 61 
 
 
 103 
 
 81 
 
 8 
 
 
 
 
 6 
 
 243 
 
 170 
 
 
 404 
 
 .118 
 
 113 
 
 88 
 
 
 158 
 
 16/ 
 
 37 
 
 21 
 
 
 104 
 
 78 
 
 81 
 13 
 
 56 
 
 in 
 
 
 137 
 9 
 
 152 
 
 1 1 
 
 1 132 
 
 1 066 
 
 1 
 
 112 
 
 1 (03 
 
 230 
 
 288 
 
 
 190 
 
 377 
 
 96 
 
 108 
 
 
 54 
 
 119 
 
 51 
 
 50 
 
 
 30 
 
 5.1 
 
 102 
 
 50 
 
 
 99 
 
 1 1 
 
 503 
 
 451 
 
 
 517 
 
 816 
 
 246 
 
 220 
 
 
 8' 
 
 116 
 
 94 
 
 98 
 
 
 139 
 
 172 
 
 12 
 
 27 
 
 
 10 
 
 21 
 
 135 
 
 92 
 
 
 178 
 
 24(1 
 
 5 
 
 10 
 
 
 9 
 
 1 1 
 
 1 719 
 
 1 (II 
 
 7 
 
 043 
 
 2 577 
 
 1 584 
 
 1 657 
 
 1 
 
 847 
 
 2 303 
 
 $12 977 
 
 $14 779 
 
 $12 
 
 128 
 
 $11 259 
 
 121 
 
 301 
 
 
 12V 
 
 299 
 
 $6 026 
 
 $13 468 
 
 12 
 
 712 
 
 $8 522 
 
 - 
 
 - 
 
 
 6 
 
 14 
 
 307 
 
 490 
 
 
 16.) 
 
 596 
 
 $1 767 
 
 $1 719 
 
 $1 
 
 698 
 
 $1 644 
 
 26 
 
 43 
 
 
 72 
 
 135 
 
 $1 265 
 
 $985 
 
 $l 
 
 765 
 
 $l 11/ 
 
 913 
 
 1 187 
 
 1 
 
 1)112 
 
 1 181 
 
 $1 886 
 
 $2 965 
 
 $l 
 
 795 
 
 $1 901 
 
 Mr. Eveleth: Yes. 
 
 Student: Have you done anything for organizations other than 
 private industry? 
 
 Mr. Eveleth: We have become involved with a couple of projects. 
 We've been active in the ABC Program which is a project for 
 making arrangements for a group of underprivileged minority 
 kids, primarily blacks, to move into a better environment, outside 
 the ghetto. ABC will typically group 20 or 25 together, and they'll 
 buy a house within a particular city or town. They're having 
 difficulty finding communities receptive to this concept, because 
 the ABC Program is financed by ABC funds for the first 2 years, 
 with the city accepting the financial obligation thereafter. They 
 have two men covering the entire United States trying to find 
 cities. What we've been able to do is to say, "Of the 40 cities 
 you're in, this is how they look. Your best programs are in these 
 25, and this is how they look. And that's the kind of city you 
 ought to look for." Instead of looking at 30,000 places, they're 
 looking at only a few specific examples. 
 
 Student: Would you comment on magazines selling lists of their 
 subscribers to direct mail advertisers? 
 
 Mr. Eveleth: A lot of the publications do this, and they make 
 some very good revenue. I can name direct mail businesses that 
 have a million customers in the United States. But if they didn't 
 rent the list, they'd lose $100,000 a year. Seriously, they run this 
 massive multimillion dollar enterprise to keep that list turning 
 over and to be able to rent it to other people. 
 
 There are two kinds of lists. There's a compiled list where 
 somebody basically takes a telephone directory or an automobile 
 registration list and puts it into the computer. All you know 
 from a compiled list is that the person is in the phone book or has 
 a registered automobile. The other kind of list is what they call a 
 buyer list or a mail order list. You get on this type of list when 
 you buy something through the mail, a magazine, something from 
 a catalog, or some tulip bulbs, etc. It doesn't matter, you're 
 on that list. These people then rent that list to another business 
 who might say, "Anybody that buys tulip bulbs would be 
 delighted with the array of accessories I have for planting tulip 
 bulbs." The list is rented, and the pamphlet is mailed to you. 
 
 There has been a campaign, called "Mail Preference 
 Service," conducted by the Direct Mail Marketing Association, 
 which is an international association. In this program the 
 consumer has the opportunity to fill out one card which is sent to 
 every direct mailer in the United States asking for the removal of 
 
 157 
 
COMMERCIAL USE OF CENSUS DATA 
 
 his or her name and address. The response, over a period of 4 
 years of public relations effort encouraging people to do this, has 
 been substantially less than 1 percent, despite the fact that the 
 association feels it has reached a potential audience of well over 
 100 million people. I n fact, out of approximately 100,000 people 
 requesting forms to have their names removed from lists, only 40 
 percent have returned their forms. People love it, that's what it 
 comes down to! The vast majority of people who get direct mail 
 advertising today love it. If that isn't true with a majority of this 
 group, keep in mind that this group, by definition, is not repre- 
 sentative of the average. 
 
 Mr. Barabba: Let me ask this question. What would happen if Mr. 
 Eveleth had been our first lecturer? What would your attitude 
 have been towards him? 
 
 Student: Most of the class probably would have been hostile, 
 because they wouldn't have understood the background or the 
 implications. I think that's been one of the most edifying things 
 about this class. We probably have greater sophistication about 
 what all this means than we did a week ago. 
 
 Mr. Barabba: Looking ahead, we have a tough decision to face at 
 the Bureau. When you finally get down to it, ask yourself, "What 
 are the questions we'll ask?" I think you could probably make a 
 pretty strong case, particularly in 1980, that anything relating to 
 energy is going to have a higher priority than pet foods. But what 
 kind of energy sources? You can't ask all of them. Which ones do 
 you handle and which ones do you leave out? In the United 
 Kingdom they ask questions of different samples so they'll end 
 up reaching maybe 30 percent of the population on a sample 
 basis. But it will be by a series of 5-, 10-, and 15-percent samples. 
 You can ask more questions that way. 
 
 I think it's going to be an interesting dialogue that takes 
 
 place between the Bureau and the Congress this time around, 
 because last time around almost 50 percent of the Congressmen 
 voted to cut the number of questions. Now we've entered into 
 the 1970's, and we've found ourselves in an era of limited 
 resources. My judgment is that the debate is going to be quite 
 different this time. I think that's going to happen, although I'm 
 not sure. 
 
 Student: On the other hand, if something is missed on the census, 
 can't it be picked up on the Current Population Survey? 
 
 Mr. Barabba: Yes, but then it's not small-area data. You have to 
 ask yourself what's the value of having this data at the small-area 
 level? In some cases, it's quite important. In the case of journey 
 to work, what's the sense of telling you that "X" number of 
 people drive to work in Los Angeles, if I can't tell you from 
 where they came? There's a report out that tells you how many 
 people drive from Downey to Burbank, and what kind of people 
 drive, how many are in the car, how many are single-passenger 
 cars, and what kind of industry they're in. To use that as an 
 example, if you just make the assumption that most of the people 
 who work in manufacturing type jobs in Downey are at Douglas 
 Aircraft and most people who are working in Burbank in 
 manufacturing are at Lockheed, you really have to ask yourself 
 why those people from Downey are driving to Lockheed, and 
 why those people from Burbank are driving to Douglas. It's not a 
 big deal when you've a lot of gas in your society, but if all of a 
 sudden you have limited resources, is it worth it for Lockheed 
 and Douglas to sit down and decide to trade employees? 
 
 Student: Will you be asking that on the decennial census? 
 
 Mr. Barabba: We did on the 1970 census. 
 
 158 
 
THE EVOLUTION OF SOCIAL INDICATORS 
 
 Lee Preston, Professor, School of Management, State University 
 of New York at Buffalo 
 
 ''Our society is very big and what the rest of society does, either on its own 
 initiative, or in response to various Federal activities, is extremely important. Studies, 
 while they may include aspects of the Federal Government, start and end with the focus 
 on society." 
 
 Lee Preston 
 
 Mr. Barabba: I first met Lee Preston through his association with 
 the American Economic Association's Advisory Committee to the 
 Census Bureau. He subsequently became the Chairman of that 
 Committee, and, as such, plays a critical role in the development 
 of its agenda relative to the Bureau and our impact on the Federal 
 statistical system. 
 
 We haven't talked too much about advisory committees and 
 their role at the Bureau, but, as I hope these seminars have 
 demonstrated, the Bureau is a very diverse organization, which 
 covers a tremendous range of subject matter areas in a great deal 
 of depth. It's a little naive to assume that any Director could keep 
 up with the varied areas that one is expected to deal with. You 
 have to have a lot of faith in the people that work for you. In an 
 attempt to review our activities, we do have a great reservoir of 
 talent available to us through our advisory committees. In a sense, 
 we make use of the advisory committees as sounding or review 
 boards. These committees play a critical role for the Bureau, and 
 Dr. Preston, in his position as Chairman, has a major respon- 
 sibility to ensure that the Bureau stays sensitive to the needs of 
 the people who rely on the data we provide. 
 
 Dr. Preston: As I understand it, and from the transcripts I have 
 read, most of the presentations have dealt with some particular 
 statistical area or type of data-gathering approach. Hopefully, 
 today, it will be useful to step away from the impact of those 
 presentations and try to see what the larger picture is about. I 
 want to take a comprehensive look at this whole collection of 
 activities and what they can be used for, and to put this all in 
 some historical perspective. Then I want to turn to some of the 
 particular applications of social indicators, which are very new 
 uses for these sets of data. 
 
 Images of Society 
 
 I think the idea behind any activity is important and it's all 
 too easy to lose the underlying idea in discussing these sorts of 
 data. In thinking about how to underscore that thought, I turned 
 to Kenneth Boulding's little book. The Image, which was 
 published some years ago. 1 If you have never read it you ought 
 to, because it is one of the most important creative works of 
 modern social science. 
 
 "Image," of course, means something in the mind, the way 
 you think about something, and Boulding says, "Image is the 
 property of the individual person," that is, it's in the individual's 
 mind. There is no image apart from that in the minds of people. 
 
 1 Kenneth Boulding, The Image, University of Michigan, Ann Arbor, 
 1956; paperback edition, 1961; quotes from pages 54 and 64. 
 
 That's what the image is. "The basic bond of any society ... is a 
 'public image,' . . . the characteristics of which are shared by the 
 individuals participating." 
 
 The way we think about society and, in fact, the existence 
 of socity as a behavioral phenomenon, means something about 
 shared images, shared ways of thinking among the people who are 
 members. Whether one thinks of the existence of a society and 
 oneself as a member, a dropout, or a member of some other 
 society, depends on the commonality of image. This com- 
 monality is a basic understanding about the way the world is. We 
 understand it together and because we do understand it together 
 that brings us into a set of social relationships that is not shared 
 by people that don't understand it that way. 
 
 This is a very general notion, but Boulding speaks at length 
 about the rise in the self-consciousness of the image of society. 
 That consciousness is the shift in modernizing and developing 
 societies from an unawareness or a lack of self-consciousness to a 
 self-consciousness. You probably are aware that in very primitive 
 societies it is quite typical to find no term at all to indicate the 
 society itself, nor a term for any other. In very primitive cultures 
 the people usually refer to themselves as "the people" or "the 
 folk" or "us" and they don't have any word for anybody else 
 because they don't think about anybody else. By not thinking 
 about anybody else, they don't really think of themselves as 
 different from anybody else. They are the whole world. It's like a 
 child's view of the world. That world is his block and he doesn't 
 really have an image of any world beyond that block. Then, as he 
 begins to develop an image of the world beyond his block, he 
 develops an image of that block existing in relation to the images 
 of that broader world. 
 
 What does all this have to do with our seminar? In the 
 whole realm of data and statistical development we are acting out 
 Boulding's notion of the increasing self-consciousness of the 
 social image. We say that we are conscious of ourselves as being in 
 society, and with that consciousness we, then, have some things 
 we want to know about ourselves which, until we were conscious 
 of ourselves, it would never have occurred to us to want to know. 
 
 The coming of consciousness-that there is a society and 
 that it has characteristics— precedes and underlies the whole data 
 development process. We don't want to collect data on the Gross 
 National Product (GNP) until we have an idea of what that is 
 Now that's true of population, too. Wc didn't start collecting data 
 on population until we had an idea of population, of a society 
 with numbers of people in it. Then the question of how many 
 people there are became a question of factual interest. 
 
 Remember, the conception— the image— comes first. The 
 image determines behavior in this broad sense, as Boulding 
 suggested. It is also important for our own immediate interest, in 
 that we establish these data-collection systems and we collect 
 
 159 
 
THE EVOLUTION OF SOCIAL INDICATORS 
 
 information. We save this, and don't save that, and assemble it in 
 a variety of ways, in response to our image of the way society is 
 and what things it's interested in, what things, as members, we are 
 interested in. It's an interactive process. 
 
 As we articulate our self-characteristics we become more 
 aware of them, and our image adapts to the process of data 
 development and we begin to think that the GNP exists— which it 
 doesn't. There is no such thing as the GNP; we made it up. We 
 invented the GNP because we had a vision of the macroeconomy 
 and wanted to get some ideas about its characteristics, dimen- 
 sions, and parts. That led us to GNP development, and now much 
 of our behavior is cast in terms of this image, GNP, which 
 numerically represents the macroeconomy. 
 
 There is an interactive process between the image, this state 
 of self-awareness, and information. We are always in it, and as our 
 society becomes more self-aware and increases its interest in itself 
 and in additional dimensions of its characteristics, we acquire 
 information about those characteristics and that information 
 refines the image. We go on evolving this image over time and it is 
 never settled, never over. It's an ongoing phenomenon. That is 
 particularly true in this important new area of social indicators, 
 because what we are doing in the social indicators field is 
 responding to the same points that were made in the discussion 
 with Dr. Jones. He emphasized very strongly that the whole GNP 
 idea, the total population idea, or any other type of single 
 statistical analysis does not measure the full range of social 
 characteristics of interest to the members of society. 
 
 What is it that we're interested in? What are the things 
 about society that matter to the members of society? These 
 define the goals and concerns that give us ideas about accounts or 
 indicators— the statistics, in other words— to be developed and 
 monitored. What is it about society or about its members that 
 you would like to know, to relate to your conceptions, your 
 goals, your concerns, your interests? 
 
 Definition of "Indicator" 
 
 This is a good time to differentiate between accounts and 
 indicators, because some literature makes a considerable point 
 about that, whereas some people say, "Nonsense, don't worry 
 about it; forget it, there's no difference." Well, both are right. But 
 it's worth making a little distinction. 
 
 I'd like to quote to you the definition of social indicators 
 out of the appendix to the book, Toward a Social Report, 2 which 
 we will refer to later. This is as good a definition as any. 
 
 A social indicator, as the term is used here, may be defined 
 as a statistic of direct normative interest which facilitates 
 judgments about the conditions of major aspects of society. 
 It is, in all instances, a direct measure of welfare and is 
 subject to the interpretation that, if it changes in the right 
 direction and other things remain equal, things have gotten 
 better or people are better off. 
 
 What to Measure 
 
 The question of refining the GNP or adapting the GNP into 
 a measure of economic welfare has been discussed. The social 
 indicators approach, which contrasts sharply with that, accepts 
 the fact that the GNP, any other economic index, or any other 
 single kind of characteristic, is not the main thing. What the social 
 indicators approach wants is a group of characteristics that are 
 very important to society. What we want is a way of getting some 
 kind of comprehensive and related reading on this group of 
 interrelated characteristics, to prevent confusing weight with 
 height. Both weight and height have to do with how big a person 
 or a table is, yet weight and height are two different things. If we 
 are really interested in sophisticated understanding of how big the 
 table is, both its weight and height are of interest. Turning to a 
 human being, both weight and height are of interest, but age and 
 income are too; we don't want to collapse those all into one 
 figure or index. We want to keep them separate, yet close 
 together, so that they can be taken in by a single view of the 
 comprehensive set of characteristics that are of interest. 
 
 If you take that viewpoint, you have joined the social 
 indicators movement. Think of the concept as flowing from 
 image to data. First, there are types of goals for social concern 
 which relate to characteristics that members of society possess 
 individually and collectively— image. Because we have these 
 characteristics in common individually, we have them as a group 
 and as a society. We're interested, for example, in whether people 
 have good health, and we're not interested in whether they have 
 blond or brown hair. We will not want to spend any time at all 
 finding out about the hair, though we may want to spend time 
 finding out about the health. We have the idea that society will 
 improve if people are healthier. We don't think that way about 
 hair, or even care whether people have any hair at all. 
 
 160 
 
 In other words, the area of concern improves or declines, 
 depending on the movement in the indicator. The indicator is an 
 output or performance statistic, not an input statistic. For 
 example, it is said, "Statistics on the number of doctors or 
 policemen could not be social indicators, whereas figures on 
 health or crime rates could be." 
 
 That occasionally numbers representing doctors and police- 
 men are used as social indicators is arguable. The reason is that 
 you can take the number of doctors as an output of society and 
 use it as an indicator. But numbers of doctors as an input is not a 
 social indicator. The budget for public health in Los Angeles 
 County is an input; it's not a social indicator as such. If you take 
 it in perspective to the income of Los Angeles County and say 
 that out of that income a certain amount is being allocated for 
 health, that puts it in an output context; it is shown as a use of 
 income. Then you could interpret it as a social indicator. 
 
 Student: When you talk about population versus policemen, 
 sometimes people measure so many policemen per thousand 
 population against the crime rate of a particular area and compare 
 those rates with another city. This probably raises a lot more 
 questions than answers. 
 
 Dr. Preston: I agree with your general point; but, you see, if 
 something is not quantifiable, then you can't say it's correlated 
 with something else, because you don't know. Then it can't be 
 the indicator. 
 
 I do not think it is wise to make an issue about the 
 difference between accounts and indicators. Most of the accounts 
 that matter, such as the GNP accounts, are subject to such diverse 
 
 2 Toward a Social Report, U.S. Department of Health, Education, and 
 Welfare, 1969. 
 
THE EVOLUTION OF SOCIAL INDICATORS 
 
 adjustment that they are essentially indicators themselves. There 
 is a difference, for example, between the GNP and the most 
 important GNP-related indicators. 
 
 Take the average number of weekly hours of work; that is 
 an indicator, it is not an account. No one cares what the average 
 number of weekly work hours is as an end in itself. We have no 
 social theory about the average number of hours worked within 
 the range varying around 40. We have sort of a norm about the 
 hours, and we have fluctuations about the norm. The fluctuations 
 are wonderful economic indicators, because one of the first ways 
 that we know that the economy is expanding is when the hours 
 of work increase, and one of the first ways we know that it's 
 contracting is when they begin to decline. The reason average 
 weekly hours is a good indicator is that it forecasts changes in the 
 GNP. But you don't know the GNP-an account— until the next 
 quarter or maybe two quarters afterwards, because you don't get 
 the data in, but you know about those hours worked every week 
 because of the payroll taxes. 
 
 Student: You're saying that an indicator is basically a subjective 
 type of thing. You just arbitrarily, through some kind of social 
 consensus, determine that 40 hours a week is an indicator? 
 
 Dr. Preston: No. The indicator is definitely an objective statistic. 
 Indicators and accounts both are statistics. However, when you 
 talk about an indicator, you are saying that you're not measuring 
 the total item you're interested in. When you talk about a health 
 indicator, you are not measuring health. You are kidding yourself 
 that this is an index of health. It is not comprehensive. It does 
 not tell you all about health, and don't think that by measuring it 
 you have measured health in the abstract. You have something 
 that when it goes up, other things being equal, health is 
 improving. When it does down, other things being equal, health is 
 deteriorating. Then you have an indicator of the change in the 
 item (health) you're interested in. 
 
 We feel there is such a close connection between the GNP, 
 our key example of an account, and the size of the macroecon- 
 omy that the GNP really is a measure of the size of the 
 macroeconomy. If the GNP goes up or down that is it. The GNP 
 is not indicating something; it is the thing itself. In the case of 
 population changes, the population is the thing itself. We might 
 have other items that indicate change in the population, say that 
 we did not measure the population directly but we were able to 
 measure changes in longevity. Other things being equal, if 
 longevity increases population is going to increase. But longevity 
 is by no means the population. That's another example of the 
 difference between an indicator and an account. 
 
 Mr. Barabba: As I recall, the definition says that when something 
 increases it is to the betterment of the objective or the concern, 
 and when it decreases it is to the detriment. 
 
 Dr. Preston: Or the other way around. Certainly your goal can 
 also be stability if that's the social consensus on the goal. It 
 depends on the way the goal and the indicator are stated. 
 
 The idea of a social consensus affords us a vision of 
 society's interests and allows us to develop the appropriate means 
 for constructing statistical data, factual data, which may even be 
 opinion data, to gain further understanding of our vision. I want 
 to emphasize that point because the word "objective" is drifting 
 into our dialogue. I am willing to stick with that, even though by 
 "subjective" one might mean attitudinal or opinion data; and that 
 
 kind of data goes into our vision as well. There is a very 
 important difference between indicators and accounts because 
 the indicator's movement makes substantial use of attitudinal and 
 opinion data, that have no place in an accounting framework. For 
 example, the questions in the new housing survey about attitudes 
 toward one's neighborhood and toward housing are very impor- 
 tant. This is a very desirable development and very much an 
 "indicators" type of development. You don't count housing that 
 way; although how people feel about housing is very important in 
 order to understand whether they are satisfied or not. 
 
 Student: If I follow your construct here, it is analogous to, 
 perhaps, the Gross National Product as an indicator with "X" 
 number of economic accounts. You are going to have a social 
 indicator with "X" number of social accounts. 
 
 Dr. Preston: No, sir. The GNP is definitely thought of as an 
 account. It is primarily an attempt to measure the size of the 
 macroeconomy and it is only because that concept is rather 
 nebulous that the GNP becomes a kind of indicator of the size of 
 the macroeconomy. It is as though you counted the population as 
 only the people over 14 because for some reason you could not 
 count all the babies. Or with the business population where you 
 count only down to a certain size and then it is very difficult to 
 count the smallest size firms because they are fugitive. They're in 
 people's garages and you can't find them. There are many things 
 that we consider hard statistics yet it turns out on examination 
 that they are really indicators right there at the edge because you 
 have the main part, but not all the information. 
 
 The other extreme is the kind of indicator where you have 
 an entirely separate item, such as the average weekly hours of 
 work. The indicator and the thing indicated are two separate 
 phenomena. For most of these non-economic social areas we 
 don't have any account aggregates to think about at the present 
 time, and I do not see any real reason to anticipate that we will. 
 We don't have an aggregate of health, although we have indicators 
 of health. All we can say is that if all these health indicators are 
 going up or down, infant mortality, longevity, days of sickness, 
 disability, and so forth, health must be getting better or worse. 
 There is not a concept of Gross National Health going from 90 to 
 95. We lack that idea. 
 
 From this conceptual level you travel to the actual 
 measurement. This is the task that you want the accounts and 
 indicators to perform. Measurement contains much of what you 
 have been doing in this seminar with data development and as 
 you have discovered there is a tremendous amount of interaction. 
 That is, we try to measure things that will be relevant to our 
 concepts, to develop statistics that will relate to our concepts of 
 society. We must employ items that can be measured fairly 
 economically and with a fair degree of accuracy. With this in 
 mind we have begun redefining some of the indicators and 
 accounts to correspond more closely with those things we can 
 actually measure. 
 
 It all goes back to our society's state of self-consciousness. 
 First, we develop the idea of "sick" versus "well." Then we have 
 the idea that "sometimes we are hotter and sometimes colder" 
 and this idea of hot and cold has some relation to "sick" and 
 "well," which suggests the indicator. Then we develop a 
 measurement to give us a precise indicator of temperature, which 
 is related to our concern about "sick" and "well." Finally, we 
 know our temperature and say, "Ah, that is our characteristic." 
 
 161 
 
THE EVOLUTION OF SOCIAL INDICATORS 
 
 Student: The Dow theory in the stock market is similar to this, 
 isn't it? 
 
 Dr. Preston: Yes, and there are lots of ways in which it turns out 
 we use the measurement to substitute for the goal, but that's 
 because the whole process is very circular, like all social processes. 
 
 Once we have the measurement, if we want to close the 
 circle, we go on to policy or action, if relevant. The effect of 
 action will be monitored by the measurement which reflects the 
 underlying goals. We'll see that incomes are up or down, 
 employment is more or less stable, or whatever it was that we 
 aimed at in our policy and action. Hopefully, we will get 
 measurements that will tell us what happened as a result of what 
 we did, and then we'll know whether our policy and actions were 
 favorable or unfavorable with respect to our concerns. 
 
 If we don't go that far, we may say, "O.K., now we know 
 about ourselves, let's just see how it goes." Then we just go on 
 living, but we also go on measuring. It's "living and measuring," 
 instead of "living and not measuring." And in this case we simply 
 find out how the characteristics we were interested in develop 
 over time, even though we may not do anything to change them. 
 This will go back to the point that a goal might be for something 
 to remain stable. Maybe our policy will be a policy of leave it 
 alone; but we'll still measure. 
 
 Mr. Barabba: These steps, even though they appear to simply 
 represent the available scientific methodology, are all fraught 
 with policy decisions. One of the best examples of this is the 
 whole concern about eliminating poverty. The responsibility may 
 be turned over to a group of technicians and they are told to 
 measure poverty, yet no one ever checks the policy implications 
 that took place in the development of that measurement. 
 
 Dr. Preston: Exactly. You go to measurement of earned income, 
 because that's the easiest thing to do. You count all the people 
 who have a certain level of earned income and then there are all 
 these people left that are below that level, so they must be in 
 poverty. Later you may find out that they're all on various 
 academic scholarships or welfare programs, or living on farms 
 where they are self-sustaining. So they really are not in poverty in 
 terms of their living situation at all. It's not to say that there are 
 not problems, or that these people are not disadvantaged; they 
 may be, but they're not living in poverty in that sense. You have 
 produced a result that is not relevant to what you had in mind at 
 all, because the measurement phase was executed in a technical, 
 "Let's just do it here in the laboratory," way instead of a full 
 scale survey. That's why the image is so important. The idea 
 underlying the whole process really matters. 
 
 History of Social Indicators 
 
 The social indicators movement fits into this process 
 reasonably well. You have had the opportunity to read in the 
 Bauer 3 book a couple of histories of the social indicators 
 movement; and in my little paper there's also another very short 
 one. I find that every one of these histories leaves some things in 
 and takes some things out. If you will bear with me, I will give 
 you one more capsule history of the social indicators movement 
 
 3 Social Indicators, Raymond Bauer, ed., MIT Press, Cambridge, Mass. 
 1966. 
 
 to hopefully put this whole thing in perspective. Then, when we 
 talk about the pieces, you won't lose the point that the pieces fit 
 together in history and they fit together through a lot of 
 interrelationships in this process. 
 
 The social indicators movement was stillborn twice. It was 
 stillborn first in the Hoover Administration in the 1929 Presi- 
 dent's Commission on Recent Social Trends, headed by 
 Wesley C. Mitchell, one of the most distinguished social scientists 
 of his day. The Commission was established under a Presiden- 
 tial charter to study and prepare a report on recent social trends. 
 It shows you the wonderful scholars they were that, in spite of 
 the fact that the world collapsed around them, they continued to 
 study the "recent social trends" based on data of various kinds, 
 through 1928. They didn't extend their thought into 1929 at all. 
 They produced a thick volume, published in 1933, called Recent 
 Social Trends, which doesn't really mention the Great Depres- 
 sion. If it had been published in 1928, which was the post date 
 that it reflects, it would have gone down in history as a major 
 contribution to American social thinking. Published in 1933, it 
 came out as a bizarre anachronism. Merely an historical document 
 it went unnoticed directly into the library archives. Of course, 
 nothing came from that exercise because the report was not even 
 filed until the end of the Hoover Administration, and it reflected 
 a conception of problems that was not appropriate for 1933. 
 
 The Social Trends Commission was followed by the TNEC, 
 the Temporary National Economic Committee, a New Deal effort 
 to study specifically the economy, but ranging into many other 
 areas of social life, as well. That Committee had a long and very 
 productive life and produced a huge set of studies which has been 
 the foundation of a great deal of modern economic research. 
 TNEC monographs, out of date now, stand as great milestones 
 and many have become incorporated into regular Census Bureau 
 series over the years. The TNEC effort might well have gone on 
 into the whole social policy analysis on a broader scale. It was 
 pointed in that direction, but World War II came along and that 
 ended that. Again, the world changed beneath the analytic 
 process underway, rendering it irrelevant to the crisis at hand. 
 
 There were no new developments until after the war. The 
 most important was the economic report which was required by 
 Congress in the Employment Act of 1946. The Employment Act 
 was one of the most important pieces of social legislation in our 
 history and the required economic report was a very significant 
 dimension of the Act. The Act said that the Executive (i.e., the 
 President) had to study the state of the whole economy— which 
 runs all the way to the edge of the entire society and into world 
 involvement— and to report and recommend to Congress specific 
 actions. This was one of the few times that the Congress ever said 
 that it took explicit responsibility for the whole of society, not 
 just particular parts. 
 
 In the 1946 Employment Act the Congress said, "We now 
 see that there is something out there called the macroeconomy. 
 We didn't know that before." It's true they didn't know that 
 before because nobody knew it. It was a new idea in the 1930's 
 that there was such a thing as the macroeconomy. By 1946 that 
 idea was well established and measurement devices had been 
 developed to the point where they could start talking about it 
 and asking for reports about specific items. Congress requested 
 such a report in the Employment Act and the whole history of 
 the economic reports has grown out of that. 
 
 The economic indicators are very closely associated with 
 the economic report. In fact, as Dr. Jones mentioned, the idea of 
 economic indicators goes back to the 19th century. Long before 
 
 162 
 
THE EVOLUTION OF SOCIAL INDICATORS 
 
 they had any idea of the macroeconomy stock market analysts, 
 investors, and bankers, were looking at numbers similar to the 
 economic reports, as clues to whether business was getting worse 
 or better. They were in the indicator stage long before they were 
 in the account stage. That's the point we are at today in most of 
 the social concern areas. We're in the indicator stage. And the 
 indicators have continued to be very important aspects of 
 economic analysis and forecasting. The Council of Economic 
 Advisers publishes a current series of economic indicators. 
 Similarly, after the HEW Department was established it began 
 two publications, "HEW Trends," and one titled "HEW Indica- 
 tors," which relate to HEW areas of concern, as the economic 
 report does to its indicators. But the HEW data were preliminary 
 and involved very tentative ideas. Those are the ideas that went 
 into the publication Toward a Social Report, which is so 
 important that we will discuss it further in a moment. 
 
 Then, there's another trend which you have in Bauer's 
 book, Social Indicators, which comes from an entirely different 
 sort of orientation. The National Aeronautics and Space Adminis- 
 tration from the outset had the idea that they should be looking 
 at the effect of NASA's programs on society. NASA was going to 
 be a vast program and also very different in terms of its 
 characteristics from previous activities. It was going to end by 
 putting people on the moon and we obviously had never done 
 that before. That might have all kinds of effects. Also, at that 
 time you could really see social reaction in the United States to 
 the Soviet space program. In fact, NASA itself was a reaction to 
 the Soviet space program. You may recall the kind of revolu- 
 tionary tremors that occurred throughout American education 
 after the Sputnik went up. We suddenly realized that some other 
 country could do something in the scientific-educational line that 
 we had not been able to do. That was the first time in a long time 
 that had happened. All kinds of revolutionary changes flowed 
 from that. 
 
 Jim Webb, who was the head of NASA in the early days, 
 realized that NASA would have unknown though vast ramifica- 
 tions; ramifications such as the fact that our response to the 
 Russian space program produced a change in American education. 
 Therefore, he wanted to set up some way of monitoring NASA's 
 impact on society. He asked Raymond Bauer, among other 
 people, to participate in designing and undertaking such a study. 
 Bauer combined this objective with a lot of his own interests, and 
 those of other people. He concluded that what they needed was 
 to begin to think about a set of basic characteristics of society, 
 areas of primary concern, and to get a reading on where society 
 was in relation to them. Then one could see how society was 
 changing over time. You had to get a base line before you could 
 make any measurement of changes. We needed to develop a set of 
 concepts and techniques that would enable us to get it started. 
 That's what they tried to do in the book. Social Indicators. They 
 didn't really succeed in doing that, but they did produce some 
 very important papers. 
 
 At that same time, another group coming out of the same 
 intellectual background produced a book called Indicators of 
 Social Change. 4 This was the work of Eleanor Sheldon and 
 W. E. Moore, at the Russell Sage Foundation. They produced a 
 very thick book which is similar to Recent Social Trends, in the 
 sense that it has chapters on a lot of topics. It's not like Bauer's 
 comprehensive, integrated approach, but the Sheldon and Moore 
 
 book is a close companion to the Bauer book. Together, these 
 represent a pattern of society's thinking to the end of the 1 960's. 
 These efforts mark the development of the measurement theme. 
 It's about what we want to know, what kinds of accounts or 
 indicators are relevant, and how they can be measured. 
 
 During the same period of the 1950's and 1960's, the 
 "goals" theme was also important. The question is: What is it we 
 desire? What are the basic areas of concern? What are the things 
 we care about? That was an issue that appeared at the end of 
 WW II, when there were clearly many efforts that could be 
 undertaken, but we were uncertain which should be done. Of 
 course, it came out of the Cold War. It's similar to primitive 
 people first being confronted with another society. The white 
 hunter comes and all of a sudden they realize that they're not the 
 only people on the planet. In that realization they come to see 
 characteristics about themselves previously unrecognized, because 
 they had never thought of there being anyone different than 
 themselves. We had the same experience in this country during 
 the Cold War, because suddenly we found that there was a 
 challenging social system on our planet that was giving us a real 
 run for our money. It was, of course, military, but it was also 
 scientific and ideological. And it was serious. That led us to a 
 tremendous attempt to return to an evaluation of our goals and 
 concerns and society. 
 
 This development led, during the Eisenhower Administra- 
 tion, to the appointment of the President's Commission on 
 National Goals, which was another group of distinguished 
 Americans. The Committee reported in a volume called Goals for 
 Americans 5 in 1960. It is not a statistical publication at all, rather 
 it concerns those things our society cares about. It's very 
 ideological and it's very American, stressing the Constitution and 
 the goal of individual freedom, the role of the individual in 
 society and the conception of society as something made up of 
 individuals rather than groups. A very individualistic and idealistic 
 conception, and a very interesting document. Despite the public 
 relations rhetoric it was intended to be a very serious statement 
 about the American people. It did not spawn any significant 
 action because it was not clear what to do about Goals for 
 Americans except affirm the value of the individual in society and 
 urge everyone to do his best. 
 
 It was followed by a more action-oriented volume. Agenda 
 for the Nation, 6 the product of another distinguished group 
 mobilized by the Brookings Institute as a guideline for the 
 presidential transition of 1968. This is a very significant docu- 
 ment because it tries to move these various goal areas into an 
 action context, an agenda for possible actions in the near future. 
 Also, the work strongly marks the transition from the Cold War 
 psychology. Of course, the notion of national defense as an 
 important goal was still very prominent in 1968, but Agenda 
 doesn't contain the Cold War visions that dominated the 1960 
 volume. These two volumes fit very closely together and are both 
 very goals oriented. They are about the image, the characteristics 
 of society that the members of society might care to know or do 
 something about. 
 
 Now the two trends have come together. The measurement 
 trend has its two main outputs in the official Social Indicators 
 volume and the Comparative Urban Indicators 7 volume which 
 
 4 Indicators of Social Change, Eleanor Sheldon and W.E. Moore, ed. 
 Russell Sage Foundation, 1968. 
 
 'Goals For Americans, Prentice-Hall, Englewood Cliffs, N.J., 1960. 
 6 Agenda For The Nation, Brookings Institute, Washington, D.C., 1968. 
 7 M. Flax, Comparative Urban Indicators, Urban Institute, Washington, 
 D.C., 1972. 
 
 163 
 
THE EVOLUTION OF SOCIAL INDICATORS 
 
 was published by the Urban Institute in Washington. The latter is 
 an outstanding piece of work, particularly for those of you who 
 are in local government. It is the first effort to bring the social 
 indicators idea down to a micro-level geographically. The meas- 
 urement trend ends there. There's nothing about goals at all. 
 Rather, here are things we are going to measure, and here are the 
 measurements. 
 
 Goals and Priorities 
 
 The goals trend continues with Toward a Social Report, the 
 HEW volume, and to a set of Brookings' studies that have been 
 published for the last 4 years. Setting National Priorities. 8 These 
 are studies of the Federal budget. What the authors have done is 
 take the base line of Agenda for the Nation and determine to 
 what extent things in the Federal budget directly or indirectly 
 relate to Agenda. Each year they examine the Federal budget in 
 terms of these categories, the goals that they think society wants, 
 and assess how the budget fulfills these various desires. It is 
 essentially a public finance, public expenditure, public program 
 analysis. 
 
 I want to explain why I don't want to talk in detail about 
 the Brookings studies. The reason is that although the authors 
 clearly say that this is not so, one would get the impression from 
 looking at the Brookings studies that the principal way society 
 accomplishes its goals and responds to its concerns is through the 
 operations of the Federal Government. They don't mean that; 
 they don't think that; but that's what these studies are about. 
 These are studies showing the connection between the Federal 
 budget and the broad areas of social concern. However, the rest 
 of our society is very big and what the rest of society does, either 
 on its own initiative or in response to various Federal activities, is 
 extremely important. Other studies, while they may include 
 aspects concerning the Federal Government, start and end with 
 their focus on society. That is the viewpoint that I would like to 
 emphasize and am interested in for my own work. 
 
 It is interesting that none of the Brookings studies are 
 Government financed. They have been generated by other groups 
 in society, although some of the people involved are, have been, 
 or will be Government people. This is not a highly partisan effort. 
 Many people feel that everything that is done at Brookings is the 
 work of the shadow Democrat Cabinet that is waiting to go back 
 across the street. There is some truth in that, but these studies are 
 not partisan studies. They are not exclusively Democrat in 
 perspective and they are not exclusively critical of the Republican 
 Administration. They are intended as cooperative, constructive, 
 analytical approaches and have been very influential. But they do 
 focus on the role of the Federal Government and make it appear 
 that everything depends on the Federal Government. 
 
 The newly established Congressional Budget Committee is 
 more or less an outgrowth of these studies. Analysis of this type 
 now can be undertaken by people in Congress who have some real 
 authority to take action when their assessments indicate some- 
 thing should be done. This approach was initiated in the first 
 Nixon economic report for 1970. The Council of Economic 
 Advisers was chaired by Dr. Paul McCracken, and at the time Dr. 
 Herbert Stein was a very important member. They put into the 
 report the idea of "national economic priorities" and tried to 
 
 "Setting National Priorities, Charles Schultz, et al., Brookings Institute, 
 Washington, D.C., 1971-1975. 
 
 make an analysis of the total GNP, not in terms of the usual 
 investment, consumption, savings approach, but in terms of "use" 
 categories, such as subsistence requirements, food, shelter, health, 
 education, and national defense. They allocated the GNP among 
 the priority uses on a statistical basis. 
 
 This was an idea that over the past two decades had grown 
 out of the National Planning Association. Stein came from that 
 organization and brought this idea into the Council of Economic 
 Advisers. It is a very important idea. It looked like it was going to 
 be the basis for a whole new breakthrough in the role of the 
 Council of Economic Advisers, to get them off the GNP and on 
 to something about what people really do in the world. 
 Unfortunately, the combination of Vietnam, Watergate, interna- 
 tional monetary problems, and inflation submerged that idea 
 entirely and it has been abandoned. I hope this idea will not be 
 totally lost. It was the most original idea to come into the 
 economic reports since they began, and certainly the first original 
 idea since 1962 when the "gap" between capacity output and 
 actual output was introduced in the first Kennedy economic 
 report. 
 
 Toward a Social Report 
 
 Let's turn now to the publication Toward a Social Report. 
 
 It is out of print. This is an important publication for anyone 
 with a real interest in the subject of social indicators because it's a 
 milestone, the first Federal publication on the subject. Also, it is 
 an outgrowth as much of a goals theme as of the measurement 
 theme, officially tying the two together in a document coming 
 from the Government, not just from private investigators. It was 
 written principally by Mancur Olsen, a professor at the University 
 of Maryland, and Alice Rivlin, who was the Assistant Secretary of 
 HEW at that time, and who coauthored the Brookings studies 
 previously mentioned. 
 
 The book's table of contents represents a very suggestive 
 list of the concerns that were identified by this group of people. 
 These are roughly the areas that you see in all the publications 
 and studies that deal with social indicators. 
 
 Looking over the table of contents one sees each category is 
 followed by a question. Under "Health and Illness" they ask, 
 "Are we becoming healthier?" That is a question about factual 
 indicators. We assume we know something about the goal. We 
 have the idea that one interesting characteristic of society is 
 whether people are healthy. Then, we immediately have the 
 question: What do we know about how well we are achieving 
 that goal? On the table of contents you have the conceptual point 
 of tying the goal areas together with measurement techniques to 
 develop a feeling for the state of society. That is the central social 
 indicator notion. 
 
 The next goal area identified is "Social Mobility"— meaning 
 primarily job opportunity and educational opportunity. To some 
 extent you can always define social mobility in a negative sense as 
 the absence of discrimination; the ability of people to move 
 freely in society in response to their interest and abilities and the 
 availability of avenues of movement like jobs and education to 
 people throughout society. This is a very general idea and 
 certainly a very hard one to measure, but you see how the 
 measurement goes when they say "opportunity." You don't have 
 to move off the farm; the question is whether you can move off 
 the farm. You don't have to move out of the ghetto; the question 
 is whether you can move out of the ghetto. Potential is really the 
 
 164 
 
THE EVOLUTION OF SOCIAL INDICATORS 
 
 question implied here. Potential is one of the most difficult 
 measurement areas but it is a very important one. 
 
 Then we have the area of the "Physical Environment" 
 which, of course, has since become, in many people's minds, the 
 number one issue. However, it was not the number one issue at 
 the time, although it was very prominent. 
 
 Then we have the traditional "Income" question going back 
 to the GNP, the distribution of wealth, family income levels, and 
 the poverty issue. With income we have a good example of how 
 we shift the definition of our goals. It is not clear that higher 
 income, or higher average income, is really the goal. It is a little 
 different from that. It is adequate income, perhaps some kind of 
 system-wide increase in income; and then we shift toward the 
 question of income distribution and elimination of poverty. It's 
 not clearly in anybody's interest for the richest "X" percent of 
 the population to get richer. You might even find that rich people 
 don't want that. Indeed, their behavior suggests it, because they 
 give money to charity, which implies that the idea of yet being 
 wealthier is not really a high priority goal to them personally. 
 Nevertheless, the income element, and the poverty question at 
 the bottom of the income element are certainly important social 
 concerns. 
 
 Next is the "Public Order and Safety" question, which is of 
 considerable interest to several people here. I'd like to refer you 
 to that chapter because it's a very interesting discussion of the 
 measurement problems as well as the transition into the attitude 
 aspects which are so important. Do you feel safe rather than are 
 you safe? I have not been mugged yet, but I thought I was going 
 to be mugged; and I did not like that. Therefore, the potential 
 dimension becomes significant, but in a way that is not quite the 
 same as with respect to income or physical environment. For 
 example, you want the air to feel clean, but from the health 
 viewpoint, you really are more concerned with whether it is 
 clean. So the fact really matters. But in the health and safety area 
 there are a lot of factors concerning attitudes that might be as, or 
 more, important than some of the outside facts themselves. 
 
 Then comes "Learning, Science and Art," which involves all 
 types of education and research. Here we are at a point where we 
 declare, "No! More is not better." That is, more years of 
 education are not better. It is not our social goal to continually 
 increase the number of years people spend in school. Many 
 people believe we may have carried that too far already. That 
 does not mean there won't be a need for additional education for 
 certain exceptional people and tasks required by the needs of our 
 society, but the notion of just perpetually increasing the years of 
 schooling is obviously foolish. What we do not know is where to 
 stop on that trend of development. Also, we do not know where 
 to shift from the input element in education, where additional 
 years of schooling are both socially and personally desirable 
 because they will benefit the person in society, to the consump- 
 tion element of education. Education as a consumption element 
 exists when additional education becomes a consumer good 
 because one likes it and one has chosen further education as one 
 might choose sports, going to the movies, or buying a bigger 
 house. Education becomes a consumption good. 
 
 The dual nature of education as a producer of human 
 capital and a consumption good gives us a dilemma as to what our 
 goal really is. On the input side, you can make a productivity 
 analysis to determine what amount of education will "pay off" 
 both socially and for the individual. That is, you want to make 
 the individual more productive for himself and for society as a 
 whole. But on the consumption side, the goal of education, like 
 any other consumer good, is to provide adequate minimum 
 
 opportunity for all. But there is no desire to have people consume 
 more education, more food, or more water, just for the sake of 
 doing it. 
 
 I think that education and research, the broad artistic- 
 scientific-learning area, is one of the most difficult indicators to 
 conceptualize. It is hard to develop a concept of what to get 
 information about which will meet the criterion that if the 
 number goes up things are getting better, if the number goes 
 down things are getting worse. Identifying numbers that have that 
 measurement characteristic and then getting the desired number 
 is very, very difficult. So much of what appears superficially to be 
 improvement, like more years of schooling, isn't. It can represent 
 a decline in the productivity of education; the education may not 
 be so good, and so it takes longer to learn the same amount. 
 That's obviously not what you want. 
 
 It's the same problem with the hospital bed indicator. Is a 
 hospital bed being full or being empty a sign of health? In a way, 
 empty beds are a positive sign, because if there is nobody in the 
 hospital they must be well. On the other hand, some kind of 
 usage ratio is a very significant sign of the availability of a health 
 facility to people who need it. Right there you have two entirely 
 different interpretations of the occupancy statistics in our 
 hospitals. The question is: How do we define and articulate goals 
 and concerns in a way that allows us to see how we are doing? 
 
 One more point about the list of concern areas: "Participa- 
 tion and Alienation." "What do we need to learn?", is not the 
 question that they really intend to ask, because it is asked all the 
 way through. The question really is, "How much social cohesion 
 is there?" Are people, in fact, involved in society; do people as 
 individuals see themselves as having a viable role as members of 
 the society instead of enemies, or alienated individuals within or 
 outside it? The discussion in that chapter is very good, although 
 the measurements are very difficult. They emphasize three 
 themes: (1) Freedom, and that's a sense of freedom. This is very 
 qualitative data, attitudinal data— whether you feel free in 
 society. (2) Equality, again, a sense of equality. It's not a 
 question about income, whether your income is higher or lower; 
 it's, "Do you feel that you are in the middle or above or below 
 vis-a-vis other people and not just in a dollar sense, but in the 
 behavioral sense of the equality of your position, and commu- 
 nity?" (3) There is the notion of Participation. Registration and 
 voting behavior are very accessible statistics and most people 
 think them relevant because they clearly relate to one's partici- 
 pation in society. If you're in society, be sure to vote. If you are 
 not in society, don't vote. That seems to give a pretty good 
 reading for an easy statistic on the cohesiveness area. 
 
 No one thinks that this is the complete list of the areas of 
 concern and social goals and we agree it is very difficult to define 
 the particular goal and its specific measures even in these areas. 
 However, on the domestic scene, these prove to be pretty good 
 catch-all areas, reorganizable in various ways for various purposes. 
 By and large, you find some version of these areas being used in 
 most of the subsequent social indicators work. In fact, you find 
 the recent Social Indicators report substantially the same as the 
 table of contents here, except for "participation and alienation" 
 which was not covered there. It could have been, with voting 
 behavior data, at least, and I imagine it will be included in some 
 subsequent study. 
 
 Discovering Social Goals 
 
 A couple of things about this list. We can always raise the 
 question of how such a list of goals is determined. There is the 
 
 165 
 
THE EVOLUTION OF SOCIAL INDICATORS 
 
 related question of how are they discovered by the people trying 
 to serve society by providing informat'on about them. Those are 
 two different questions. The "How is it determined?" question is 
 very hard. You can say, "Well, it's determined by all the processes 
 that take place in society. Teaching, home upbringing, social 
 forces, catastrophies, leadership, and everthing else, go into 
 determining a person's view about the way the world is." 
 
 At the theoretical end of the social indicators field, where 
 considerable work is underway, people are trying to develop a 
 conceptual model of society that will have some of these 
 dimensions. You start with a theory about people, and then 
 theorize about them interacting, which is, of course, society. Out 
 of that theory about people and their interaction will come some 
 performance dimensions which will be the areas of social concern. 
 That is a very ambitious undertaking and some progress is being 
 made. 
 
 I think that people at the action level, more like ourselves, 
 do better just to accept the notion that there are some things that 
 the members of society are interested in and some things that 
 they are not. We would do better to shift, therefore, from the 
 question of "How is it determined?" to the question of "How is 
 it discovered?" If we are interested in fulfilling a role here, we 
 would like to be able to observe for ourselves and to report to 
 society information relevant to the areas that society is concerned 
 about. Therefore, you shift out of this sort of theoretical notion 
 and into an empirical notion of determining what it is people are 
 concerned about. Very sophisticated techniques about this are 
 available and should be used to find out what people desire. 
 
 A good example of the problem of determining what 
 people want has to do with health, particularly the Government- 
 supported health insurance question. As you know, this has been 
 a subject of great discussion and experimentation for at least two 
 decades in this country, and nowhere more so than in California. 
 Throughout the 1960's the question was: How much do people 
 really care about that? That is, what do they really want in the 
 way of health care protection? 
 
 Many studies have been undertaken to determine that. Two 
 studies that I have were undertaken at almost the same time. One 
 was a so-called Delphi technique study, a "polling the experts" 
 study seeking the views of the best-informed people. The other 
 was simply a random sample of public opinion. Both were aimed 
 at discovering very broad areas of social concern. They were 
 conducted by The Conference Board, formerly the National 
 Industrial Conference Board. Vietnam dominated both lists. 
 Setting that and the international items aside, the two lists were 
 very similar, which is reassuring. Both the experts and the random 
 sample of the larger society indicated the same rankings and the 
 same kind of priority list, with one exception— health. The 
 experts just didn't see it! It didn't have any prominence on the 
 experts' list, apparently because they come from a part of society 
 in which that is not a pressing personal problem. The experts 
 about society did not recognize how widespread the feeling 
 favoring some kind of health care protection was. The random 
 survey showed health care as the highest priority. Since then 
 many polls and surveys have shown this public concern and 
 Congress, by its behavior, has recognized it. President Ford, in his 
 first address, said the one thing he wanted for certain was the 
 establishment of this health care program. He's obviously trying 
 to do what he thinks is at this time the appropriate thing for the 
 needs of the country as expressed by the public. This whole story 
 is a very good warning to all who work in this area and are apt to 
 
 think that experts are aware of what society wants. You see, the 
 experts didn't know; you have to ask the people. 
 
 Mr. Barabba: Let me carry that one step further because the 
 period of your two surveys was a time when my former company, 
 DMI, was doing research in this area. As I recall, Dick Wirthlin 
 mentioned to the class the "Alias Smith and Jones" technique 
 where, rather than identify the institution that held a point of 
 view and then measure the public's attitude towards the 
 institution, we put a hypothetical person in place of the 
 institution. That, basically, was a concept designed for a study 
 dealing with the measurement of public attitudes toward two 
 national health insurance programs as envisioned by Senator 
 Kennedy and the American Medical Association. You can all 
 appreciate the differences that you would be measuring if you 
 said the AMA feels this way and Senator Kennedy feels another. 
 That's why we came up with the "Alias Smith and Jones" 
 technique. 
 
 The example that I remember best was Montgomery 
 County, Ohio, which is the Dayton area. It's an interesting, 
 completely self-contained community. It tends to have a two to 
 one Democrat registration and is very heavily unionized. We 
 conducted a study there and before we gave the results to the 
 local physicians, we gave them the questions and asked them how 
 they believed the public would respond to the two views. The 
 Montgomery County Medical Society, almost to a man, predicted 
 the community would identify more with Mr. Smith (Senator 
 Kennedy) than Mr. Jones. Yet it was just the opposite in the 
 community! The community identified more with the American 
 Medical Association, favoring less involvement in the program by 
 the Federal Government and more participation by private 
 insurance companies and medical societies. 
 
 The doctors did not believe the study. They argued that the 
 study did not reflect what the newspaper headlines said— that 
 society needs more and more of this. We found that we really had 
 to convince the doctors by actually going through each step of 
 the study to prove that it was a systematically drawn sample, and 
 the procedure was as unbiased as possible. 
 
 We, then, duplicated that example with a leadership 
 meeting of the American Medical Association where they had 
 doctors from all over the country. We theorized an area like 
 Montgomery County, asked them the same questions, and the 
 exact same phenomenon took place. They saw the public as really 
 being more against the American Medical Association than it 
 actually measured to be, further indicating experts are not always 
 aware of society's desires. 
 
 Student: In your Delphi approach, were your experts all physi- 
 cians? 
 
 Dr. Preston: No. This was a nationwide social concerns approach. 
 I don't know how many of them were physicians, but there 
 wouldn't have been very many. This was an attempt to find out 
 what the major areas of concern to society were perceived to be 
 by a broad cross section of experts. 
 
 Student: I was wondering if your approach would be slanted by 
 who you used as your experts? 
 
 Mr. Barabba: I think in any kind of a Delphi technique that's an 
 inherent problem. 
 
 166 
 
THE EVOLUTION OF SOCIAL INDICATORS 
 
 Dr. Preston: This is a very significant point. In designing this 
 social indicators report it would be very easy for a group of 
 people to sit down and decide what the major categories are, 
 gather the appropriate statistics related to the categories and 
 produce a beautiful, fully documented report. But if you gave it 
 to the man on the street and said, "These are the areas that are of 
 concern to society," he'd say, "They're not areas I'm concerned 
 about." If one is really interested in developing material of this 
 sort for social decisionmaking purposes, we'd like to get the man 
 in the street's views into this process as early and accurately as 
 possible. This will help the process generate the self-conscious 
 image of our society that reflects those concerns of society. 
 
 There's a role for explanation and leadership in this process 
 that is important. It's like the courts. When we have a real jury 
 trial with the adversary positions and evidence presented, the aim 
 is to get a consensus from the jury. The consensus of the jury is 
 the judgement of guilt and innocence. In that same sense society 
 determines its social goals. 
 
 Like the jury trial, there's a role for advocacy in presenting 
 evidence— the reasons for a goal, the justifications, what it really 
 means, the cost, the effect, and those types of things. You need 
 to have all sides presented if you are going to get a consensus 
 back that is worth anything, that is really accurate. 
 
 The whole social indicators' focus is on performance 
 attributes. We ask: What is the output of the system? More than 
 that, what is the characteristic of society that is related to the 
 output of this system and how is that getting along? Not just 
 having more people in the hospital beds, but actually more well 
 people out on the street. That is health, and that's the thing you 
 want to focus on. 
 
 I'd like to shift to the National Planning Association study, 
 headed by Dr. Nestor Terleckyj. This study first identified areas 
 contained in Toward A Social Report and other general literature, 
 and one or more indicators to go with each area on a 
 measurability basis. They only chose things on which they could 
 get information believed reliable and not overlapping. Then, using 
 the decade from 1972 to 1981, they tried to forecast the trend 
 for each of these indicators within the defined goal areas. In other 
 words, they took a goal area and anticipated simply from the way 
 things are going; if no changes occur, the indicator will display 
 some pattern of change over the future. Then, and this is the hard 
 part, they go to the problem of estimating feasible changes from 
 the trend. This involves a combination of technical and economic 
 analysis. For example, "X" number of people have gone to the 
 moon and if that trend continues for the next decade, "X" many 
 more people would go there over that period. Then, on the basis 
 of technological feasibility, it's theoretically possible the whole 
 population could be sent to the moon. However, that is absurd. 
 Conceivably we can send 2 dozen or 4 dozen persons at a certain 
 cost within the realm of possibility, if we choose to do that. The 
 alternative to the projected trend is the difference based on a 
 technically and economically feasible program. 
 
 You have a range of choices between where you'll be if you 
 don't do anything different and a sort of maximum conceivable 
 change. They break the decade down into two periods because in 
 so many of these things the timing is critical. You would have to 
 do so much within the first 5 years if you were going to be able 
 to do so much in the next 5 years. 
 
 They add one more dimension— and this is really a hard 
 dimension— they ask what a program would cost. You take the 
 same time chart, except you have cost instead of the indicator. 
 What would it cost in terms of resources to achieve the estimate 
 
 as opposed to other programs? What would it cost over what you 
 would spend on that program otherwise? Again, in the cost you 
 have a starting point, the cost which you are already experi- 
 encing; and you would anticipate some costs continuing into the 
 future, no matter what. If you wanted to go up to the maximum 
 possibility, what would the cost be then? They break that down 
 into time periods, which is so important because the total cost 
 implications will be very different depending on the timing 
 factor. To get the total cost you have to take the cost per year for 
 the whole decade and timing has a tremendous impact upon that. 
 
 You can imagine how difficult this work is and how many 
 guesses, estimates, and adjustments are involved. The methodol- 
 ogy is important because it represents a big step beyond existing 
 social indicator work, by going on, first, to the trend projection, 
 recognizing that the indicator can be expected to change over 
 time merely as a result of society going on; then, estimating the 
 maximum change; and, then, tying that to the cost and the timing. 
 Those all are new conceptual additions, and very, very significant 
 ones. 
 
 Another virtue of this approach is that it shows the 
 interrelation of goals. You start with society as a whole, with all 
 its characteristics which are interrelated. Then you begin to 
 identify the goals, and that provides separation. Then you 
 identify the indicators and that's very separate, because they have 
 to be specific. Then you go on to the activities and policies, and, 
 again, they are concrete; that is, whether an activity is going to be 
 in HEW, at HUD, or the police department. Where is this activity 
 going to take place? But the impacts, the results of those policies, 
 are going to manifest themselves within the society at large. 
 Therefore, the analysis helps you bring back the interconnection 
 within the system, which is where you started and where you 
 want to finish. 
 
 It's hard to end there when you go into increasingly 
 particular types of items. We all know that the social programs of 
 the 1960's embodied, as one of their characteristics, an attempt 
 to have one program solve one problem. We had programs that 
 were supposed to help correct one problem, and perhaps they 
 did, but over in an adjacent problem area terrible consequences 
 ensued. This NPA analysis is designed, for all of its shortcom- 
 ings, to help you see these side effects. It's designed to bring out 
 the interconnections, which one doesn't want to lose in any kind 
 of comprehensive social analysis. 
 
 The other point about this type of analysis is that it gets 
 you right up to the edge of trade-offs among the activities and, 
 therefore, among the impacts. Terleckyj does not tell you which 
 one of these to do. He goes into this cost analysis and then says: 
 "If you add up all these costs and all the resources in society, 
 these costs will outrun the resources; so you have some very 
 difficult decisions to make." And that's the end of his presenta- 
 tion. He's not trying to decide for you because he doesn't 
 consider that his role. I think that anybody with a political 
 feeling for the notion of letting society make decisions for 
 themselves would say that he's right. We want to know what we 
 can decide, and then bring the decision itself back to the political 
 process, back to public opinion and voting. 
 
 This is a type of input information for social decision- 
 making. It says that you could have a maximum impact in this 
 area with "X" dollars, or in another area for "Y" dollars. You 
 may not have enough dollars to do both, or you may not want to 
 use enough dollars to do both. You m«y not want to give up 
 something else to get the dollars to do both. There is your choice. 
 It includes the trade-off idea as well as the interrelatedness idea. 
 
 167 
 
THE EVOLUTION OF SOCIAL INDICATORS 
 
 Interrelatedness means you can take one activity and have lots of 
 different impacts. Trade-off means that with a number of 
 different goals you may have to make a choice among activities, 
 since you can't have them all. This way of thinking puts the key 
 social trade-off problems into perspective, and permits one to 
 think about them in output terms. 
 
 Student: But, Terleckyj's work is really only a first attempt? 
 
 Dr. Preston: Yes. It is very significant as a first attempt. No one 
 would imagine that Terleckyj and his associates could really 
 produce factual answers to these difficult questions on the first 
 try. But the idea is extremely significant. 
 
 There's one other thing about it that I want to call your 
 attention to. When Terleckyj talks about cost, he shows total cost 
 figures, and these are the cost increase figures, not the cost of the 
 activity in a budget sense. Costs are broken down, first, by time 
 (which is the point I emphasized already), and, second, by public 
 and private. He's emphasizing here that given the present 
 structure of our society, which we anticipate to continue into the 
 future, we would expect that all of these social endeavors would 
 involve some combinations of public and private activity, and the 
 appropriate combination has a lot to do with which of the 
 activities are chosen. The size of the public sector is of critical 
 importance and is partly determined by which of these activities 
 are to be the social choices. If we choose activities that are in the 
 public sector, like national defense, we won't expect to have a big 
 private endeavor. If we choose other activities, they may all be 
 private, or they may be a combination of public and private. 
 
 It's the structure of this analysis that's important, because 
 this type of analysis can be carried to many particular areas— to 
 particular localities, to important goal areas. This work represents 
 a technique of conceptualizing the comprehensive impact of the 
 way the public and private sectors do business. I think that it will 
 not be too long, with the new budget committee and as a result 
 of the Brookings study, before some of this framework finds its 
 way into official documents and becomes part of the image, part 
 of the way we perceive the system working. Remember, our 
 perceptions have so much to do with the way the system really 
 does work. 
 
 Student: The difficulty is getting consensus on the goals. 
 
 Dr. Preston: That's not quite so. If we go back to Toward a Social 
 Report, the table of contents, I don't think there's much 
 confusion about the general direction of desired development in 
 our society. I think that we have a much stronger social consensus 
 about goals than most people realize, and all we have to do is 
 bring forward some concrete proposals in many of these areas. 
 Health insurance is an example, where all we have to do is get a 
 concrete proposal before the political decisionmaking system. We 
 have to get the proposal out there to find out what the consensus 
 is. We can't get the jury's consensus before it goes in the jury 
 room and votes on the evidence presented. They may have a 
 perfect consensus sitting there in the jury box, but we can't say. 
 
 Student: The problem would not be arriving at the goals. The 
 problem would be correlating the goals with our free enterprise 
 system. I think that's where the hold up will come. 
 
 Dr. Preston: That's great! That's the next step! I agree with you 
 100 percent! I think there's a very broad consensus about goals. 
 
 168 
 
 But I think there's a lot of uncertainty, more than disagreement, 
 about how to act to achieve the goals. We want health standards 
 to be better. We want better access to economic and social 
 opportunity. How will they be realized? It is not that we disagree 
 about having it, rather, we don't know how best to accomplish 
 our goals. Fortunately, we're wisely distrustful of any one 
 solution. We are pretty sure there is not just one solution, so we 
 need to look for systems that will allow us to experiment with 
 many answers, and let different answers solve different problems. 
 
 It's like curing unemployment. The easy way to cure 
 unemployment is to shoot the unemployed. That is an easily 
 overlooked point. If it is really true that getting rid of 
 unemployment is your number one goal, stand them up and 
 shoot them. Of course, the alternative is to put them in the army 
 and send them somewhere else to get shot, which has been used 
 historically. There are lots of easy solutions but it turns out that 
 we don't want those. Our problem is to see how to use a 
 multiplicity of solutions. And that's where the problem is. Not 
 the consensus about big goals. I think we are a strongly 
 consensused society. 
 
 There was a consensus during the 1950's, the Eisenhower 
 years, in our society. Then in the 1960's came a period of 
 division. I think we're back in a period of consensus, but the 
 substantive elements of the concensus are not the same as in the 
 1950's. We have new dimensions about which there is agreement. 
 Now we are trying to articulate our actions and policies on the 
 basis of this new consensus. 
 
 Student: If this is the framework for measuring these goals, then 
 would it follow that you could take these as an aggregate and 
 come up with an index for the quality of life? 
 
 Dr. Preston: I'm very distrustful of bunching together dissimilar 
 elements. Dr. Jones talked about the Consumer Price Index. We 
 want to know the cost of living and what we have is the CPI. The 
 CPI is a fine statistic but it isn't the cost of living. We don't want 
 to discontinue it because we're hung on it, it's part of the image, 
 and we don't have anything to substitute for it. You can't beat 
 something with nothing. But we need an alternative. 
 
 Let me give you another example of combining everything 
 in an index. In the early 1960's, Professor Aaron Gordon, who is 
 presently the President of the American Economic Association 
 and a leading economist, with special interest in unemployment, 
 was appointed by President Kennedy to review the unemploy- 
 ment statistics. There was a question about the accuracy and the 
 appropriateness of the unemployment statistics, like there is 
 about the CPI. He studied it, with a group, and within his 
 committee he alone recommended the discontinuation of the 
 global unemployment rate, that is, the national unemployment 
 percentage figure that you see in the newspaper. He said that he 
 thought that the global rate was a bad indicator, even though the 
 components were fine. You weigh the components together and 
 produce a global figure which could mean anything. It depends 
 upon the age structure of the population, the flow of young 
 people into the labor force, retirement rates, and several other 
 variables, so that anything could be happening as to the true facts 
 about unemployment yet there could be the same changes, or 
 lack of change, in the national unemployment rate. 
 
 Professor Gordon tried militantly to get that unemploy- 
 ment rate eliminated, but the other economists and the bureauc- 
 racy wouldn't stand for it. It's like the fixation with the GNP. 
 The bureaucracy had too much of its image tied to the national 
 
THE EVOLUTION OF SOCIAL INDICATORS 
 
 unemployment rate; people had spent all this time lecturing the 
 Congress about the importance of the unemployment rate. They 
 said, "Oh, if you discontinue it the whole lesson will be ruined." 
 So they didn't do it. 
 
 I don't care about that statistic particularly, but I agree 
 with Gordon's basic idea, that you want to stay closer to reality, 
 if you can, without getting a thousand indicators. That's too 
 many indicators, but a hundred is not too many, and 10 per goal 
 area is not too many. Not at this stage of knowledge. Ten per area 
 is much better than one. But, the notion of putting it all together 
 into an index of well-being is not something I would agree with. 
 
 You discussed with Dr. Jones the Measure of Economic 
 Welfare (MEW) -changing the GNP into the MEW. He opposed 
 that and I certainly do, too. The GNP is fine for what it is. If you 
 need something else, you need something else, but don't try to 
 change the GNP. You shouldn't put frosting on your steak 
 because you like dessert; why not have both? They are two 
 different things. I strongly favor disaggregation, keeping the 
 indicators separate, and keeping the goal areas separate, so that 
 you can view the effect of different policies. 
 
 Mr. Barabba: To give you an example, take the Wholesale Price 
 Index. That number came out and people said, 'There's going to 
 be an economic disaster because it increased by such a large rate." 
 When we look at its components we see that several reasonable 
 actions have taken place, which increased the whole average for 
 that particular quarter. Does the average affect the whole society 
 in its economic well-being, or can we compensate or design a 
 program for those areas that increase? What we hear about is that 
 the Wholesale Price Index went up. What should be talked about 
 is what components of the Wholesale Price I ndex went up. 
 
 Dr. Preston: You really can't make any interpretation of the 
 increase in the total index without going to the components, 
 which may well mean the components should be the object of 
 any action. 
 
 Mr. Barabba: Today we hear that the unemployment rate is 
 holding steady. It is, yet there's so much evidence of a 
 tremendous amount of turmoil in our society, but we're not 
 talking about that because we're so locked in on something called 
 the unemployment rate. 
 
 Local Area Data 
 
 Dr. Preston: I want to make one more shift and focus on local 
 areas: bringing down the social indicators analysis to cities. 
 States, counties, metropolitan areas, or whatever the so-called 
 small area is. Of course, a lot of them are not small at all. I want 
 to emphasize this direction of development because this is the last 
 item we need to cover before we go into the use of social 
 
 indicators in corporate management, which, at the present time, 
 is the area of major activity in the social indicators field as far as 
 empirical work is concerned. The other would be the theoretical 
 work that I talked about earlier; the development of a theory of 
 society that would have performance attributes that could be 
 measured by the indicators. The empirical side, which reached a 
 high point in the publication of Federal data on social indicators, 
 is now pushing out to study urban areas. States, and other types 
 of subnational groupings. 
 
 The Urban Institute set out first to make a study of social 
 indicators for Washington, D.C. In a preliminary document, they 
 discovered, as always, that the data yield a number for something 
 and you necessarily ask whether it is high or a low number. 
 Since you have just one number, you don't know. You need a 
 base for comparison, a base for interpretation. Therefore, they 
 pushed into a time dimension saying, "We'll consider the trend. 
 What has it been, what will it be?" That's an easy thing to do. 
 They decided to consider intra-regional patterns, such as city/ 
 suburbs, which is one of the things frequently studied in these 
 various preliminary reports. Some of the studies, like the one for 
 Charlotte, N.C. has a black/white comparison, which is extremely 
 relevant for many cities, in order to make a disparity analysis in 
 that dimension. There are other dimensions that could be used. 
 
 In Washington they needed a larger framewoi k of compari- 
 son so they took 18 cities. In order to work on an 18-city basis, 
 they simplified their analysis to one indicator for each goal area. 
 They could get by with one indicator per goal area, in part, 
 because they had a rather large number of goal areas, as compared 
 to some of the studies that have just a few very general areas and 
 many indicators within them. They had a larger, more detailed 
 list of quality categories and they had one indicator for each. 
 That led them to generate a lot of tables that permitted the 
 comparison of indicators for each area across this group of cities 
 and composite rankings. They made most of their analysis in 
 terms of rankings. It's a simple comparison, and it isn't totally 
 wrong if you are a little careful about how "finely" you rank. 
 
 They focused primarily on Washington to see whether it 
 was improving faster, at the same level, or deteriorating faster 
 compared to the 18-city average. They looked to see whether it 
 was high or low with respect to the 18-city average; they looked 
 at the disparity measures. They constructed this basis for 
 comparison by looking at the other places that they thought 
 presented at least a relevant analogy to the Washington area. They 
 present the raw data for the indicators and they present rankings 
 for the indicators. 
 
 This document has set off a whole set of similar, related, 
 types of studies in all of the metropolitan areas of the United 
 States, in one form or another, under one sponsorship or another. 
 I think this is extremely appropriate. The global aggregates, on a 
 national basis, are significant as reference points, but that's all 
 they are. The policy initiatives, improvements, and changes have 
 got to be observed and carried out at local levels. 
 
 169 
 
SOCIAL INDICATORS FOR CORPORATE MANAGEMENT 
 
 Lee Preston, School of Management, State University of New York at Buffalo 
 
 "If we are looking for a system that will enable the managerial units to continue to 
 function through time, we need a system similar to market forecasting that will not 
 simply produce a list of current goals and desires but rather new information about what 
 people will want in the future." 
 
 Lee Preston 
 
 Dr. Preston: All corporate organizations, which are the mana- 
 gerial organizations in our society, are set up for some social 
 purpose. They don't have rights, as humans do, to do whatever 
 they please; and there's no reason that they should. The approach 
 that we have historically taken, that corporations can do 
 whatever they please within their charters, is fundamentally 
 unsound; we are about to abandon that idea. I think that's a 
 factual observation. The extent within our society of various 
 kinds of regulations— for example, the requirement of cooperating 
 with census studies, demanding even confidential data— is already 
 a statement that society requires many corporate acts performed 
 without any choice in the matter. Add that to the tax structure. 
 In a situation where half of the formal corporate profits go 
 towards taxes, who can say that there's any limit to the power of 
 society over these entities? There really isn't, and I think we're 
 realizing that now. Therefore, we're looking for the implications 
 of that realization. 
 
 When I talk to business groups the title that I usually use 
 with this material is "Socializing the Corporation." I intend that 
 to be a title that will cause people to at least listen through the 
 first few minutes to see what on earth this flaming radical is going 
 to say. I find many people say, "Don't we already have 
 socialism?" I have to answer, "No." But I think their question 
 reflects an awareness of the fact that we are operating in a 
 self-conscious society that is deliberately setting up numerous 
 goals and systems for the achievement of a variety of social ends. 
 
 If for no other reason, I do not think that we have socialism 
 because of the relative lack of centralization of decisionmaking. 
 In spite of the tremendous role of the Federal Government, the 
 truth is there are very few decisions within our society that are 
 made and implemented at a centralized level. There are decisions 
 made at the Federal Government level, and they have vast 
 implications, but it's not like the Army. It's not a set of decisions 
 made and then executed down the line through the rest of 
 society. I presume we want to continue that unstructured method 
 for ideological reasons and also, I believe, for efficiency, because 
 centralized decisionmaking is very inefficient, because it involves 
 so much standardization. Standardization means that the solu- 
 tions don't fit the problems since problems by their very nature 
 require flexibility. 
 
 In response to the comment this morning about the 
 problem of integrating the private sector into this kind of social 
 consciousness, I think we are groping for ways that will permit 
 the private sector to maintain the kind of independence, 
 initiative, and responsibility that has characterized it, while 
 adding these new social concerns. To find these ways we are 
 taking the same kind of ad hoc, pragmatic, "Let's try this, let's 
 try that" approach that we typically employ in the private 
 
 economy, and adding a response to many of these other areas 
 about which society is clearly interested. 
 
 The alternative to that, of course, is regulation. We see a 
 great deal of regulation and it is increasing. Some regulation is 
 desirable or necessary if we find that we can't get people to do 
 anything any other way. As people in the private sector or other 
 organizational managers come to see themselves as operating in a 
 social context rather than in terms of any single variable such as 
 profits, we can expect them to be looking for managerial 
 techniques that allow the integration of that social context into 
 their management. 
 
 This is not really anything new. It's not as if for the first 
 time it were being proposed that organizations have more than 
 one goal. Organizations have always had more than one goal, and 
 people who say that in the end it all comes down to profit are 
 mistaken. 
 
 You say, "What about the market share?" Oh, well, it 
 comes out of profit in market share. "What about growth?" Well, 
 yes, growth. "What about the stability in the organization?" 
 
 Usually you can increase profits by discharging all the top 
 executives, since they're all overpaid. The answer is that we have 
 to maintain the continuity of the organization and, therefore, 
 must have prizes to keep people working through the system. 
 Don't we have air-conditioning even though there's no law 
 requiring it? Then you have company recreational facilities, 
 whatever they may be; the flashy annual report that has no 
 connection with anything, but which is a very expensive 
 undertaking; and the big new building that is a great landmark 
 downtown. It will be years before the building turns a dime, 
 though it establishes us as one of the leading firms in the 
 community. Yet, some say that it all comes down to profits. I 
 argue that it's the image. Firms really have images of themselves 
 that are quite different from "it all comes down to profits," 
 although most firms don't yet have a substitute ideology. 
 
 I see the corporate sector looking for a different way to 
 perceive its role in the system, its diverse connections with 
 society, rather than saying, "Well, our real religion is profits, but 
 we just do all these other things as extras." Frankly, it looks like 
 the extras are pretty big and profits not so big, and it seems they 
 have a different religion than the critics say. I feel that this is a 
 phenomenon well under way in the corporate system. That's 
 what I call a socialization process. "Socializing the corporation" 
 is not nationalizing the corporation. It is bringing the corpora- 
 tion, and its management, into an awareness of its existence in 
 the social context rather than having it think of itself as primarily 
 an economic entity with the primary task of profitability. This is 
 not to say that the economic profit goal will be suppressed. It's 
 there, but supplemented by all these other goals. 
 
 170 
 
SOCIAL INDICATORS FOR CORPORATE MANAGEMENT 
 
 Mr. Barabba: Along these lines, in past ages of corporate 
 concerns, you had cases where a firm dominated a community. 
 They were actually criticized as patronizing the workers for 
 providing, in a sense, welfare services. The Hershey Company in 
 Pennsylvania was severely criticized for building a company town 
 and a place where everything was provided. Corning Glass is a 
 good example of a company that contributes significantly to the 
 cultural well-being of the community. General Motors in Flint, 
 Mich., provides the whole recreation program of that city. You 
 do have tangible manifestations of a corporate social conscience. 
 
 Dr. Preston: Very definitely. The company towns present the 
 whole spectrum. There were company towns that were simply 
 idyllic. There are company towns, even today, that people will 
 not leave in spite of promotions because they feel it's just a 
 wonderful place to live with all these facilities provided. It's 
 typically a smaller place where you have all the amenities of a 
 large city because it's a profitable company town. Of course, the 
 other extreme is the archetype of the company town where it is a 
 huge sweatshop and everything is taken back before it is paid out. 
 
 An individual was telling me about a town where his 
 company operated. Historically, the company had been very 
 benevolent in a moral sense, and intentionally so. It was not a 
 town that was fraught with difficulties; it was a very harmonious 
 community. The company decided to try to withdraw as much as 
 it could from the company town aspect. One of the main 
 elements of that withdrawal was to sell off the houses so that the 
 indigenous population would become homeowners. They would 
 then be the residents of the town on an ownership basis. It would 
 be an ongoing community without regard for the company. 
 However, when the houses were sold the improvement in the 
 residences, property, and the community involvement of people 
 in the town was simply incredible! The company executives could 
 not believe it! They had thought that everything was really in 
 good shape. But, the ownership of the houses and the property in 
 general changed the entire social atmosphere. Voting and com- 
 munity organization changed very dramatically, with much more 
 participation. This indicates that there are some behavioral 
 elements that you don't see in economic, statistical, or numerical 
 ways, but are there to be observed nonetheless. 
 
 This corporate sector movement, this awareness, this 
 socialization process, has been underway for at least a decade. 
 The idea itself is really 20 years old, and it dates from a 1953 
 publication called The Social Responsibilities of Businessmen. 1 
 This publication had little or no impact because those were years 
 of stability; at best it was viewed as a benign contribution. 
 
 The student and civil rights revolutions in the 1960's 
 brought all this to a head. The previously self-satisfied corporate 
 executives found themselves being bombed on their heads, first 
 by people on the street and then by members of their own 
 families. They realized that something was out of tune, and that 
 they had to pay attention. There was a crisis response by all the 
 large businesses in the United States, and I know that you 
 observed this in your own environments, wherever you were at 
 that time. Certainly here in California you observed it specifically 
 with Bank of America. 
 
 A lot of those stories had a beginning, a middle, and an end. 
 They appeared then to be ad hoc phenomena; but clearly they 
 were not. They were diverse, and each one was a unique case. 
 
 1 Howard R. Bowen, The Social Responsibilities of Businessmen, 
 
 Holt, New York, 1953. 
 
 These incidents resulted in this very different view of the 
 relationship between the large organization and its environment. 
 This naturally raises questions about how the relationship is to be 
 monitored and managed. How is the relationship to be under- 
 stood by the managers, by those people who are responsible for 
 controlling these organizations? It is absolutely ridiculous to 
 think that they can just give away all the capital of the firm to 
 good causes and then close up. They could do it if they wanted to 
 and if they were so licensed, but that would not accomplish the 
 tasks that the society has indicated it wants these organizations to 
 perform. Society has indicated its desires by buying their 
 products. Theoretically, if you don't want a firm in the market, 
 then just don't buy its product and it will go away. There are 
 some monopoly cases where we find it very difficult to stifle 
 undesirable firms. Satisfying material desire is the function of the 
 consumer goods firm and similarly in the public sector we have 
 these organizations to provide health care, police service, water, 
 electric lights, or banking services that are vital parts of the 
 system. 
 
 On the other hand, it is quite clear that corporations affect 
 people's lives, neighborhoods, incomes, and everything about the 
 rest of society to greater or lesser degrees. If management is going 
 to bring that awareness of these consequential effects of the 
 primary activity into its process, the question is how to manage in 
 this "social consequences" atmosphere instead of in the "balance 
 sheet" atmosphere and without throwing away the balance sheet? 
 
 Criticisms of Social Involvement 
 
 The notion that corporations can solve all the problems of 
 society is acknowledged as ridiculous. The criticisms from those I 
 call the "Fundamentalists"-the people for whom Milton Fried- 
 man is the spokesman— are very valid as regards the hysterical 
 crisis kind of "General Motors, save the country!" screaming. 
 Briefly, he makes three points in his classic criticism of this whole 
 approach. One of them is that the job of General Motors is to 
 make cars, or to do whatever else it's supposed to do, and 
 anything it does that causes it to be less efficient is a deviation 
 from that function. 
 
 Now, that's true. However, the answer to that is that GM 
 doesn't just make cars, it also employs people, affects neighbor- 
 hoods, does this and this and this. Therefore, even though its 
 reason for existence is to make cars, its consequences are these 
 other impacts. It's not a question about whether you want these 
 other impacts or not, they exist. A management that doesn't 
 acknowledge these impacts, or that there are ways society wants 
 them to behave, would be as ineffective as a management ignoring 
 the things society wants in its cars. It's like shooting all the 
 unemployed or discharging all the executives to save money. 
 
 The second fundamentalist point is the very serious 
 question of competence. By what competence do these corporate 
 managers know what would be good for society? They're not 
 trained in that. There's nothing in the corporate management 
 technique that gives them an idea about what would be socially 
 beneficial. We have seen in the last decade quite a number of 
 well-intentioned corporate ventures which were terrible. Others 
 certainly caused less good than they might have because the 
 people that tried to do it didn't know what to do. 
 
 That's true. If it's a function the business must perform you 
 add competence in that function, just as you do in engineering, 
 accounting, marketing, or whatever else it is you need in order to 
 perform. 
 
 171 
 
SOCIAL INDICATORS FOR CORPORATE MANAGEMENT 
 
 The third issue is legitimacy. The Friedmanites say corpo- 
 rate activities affect society at large. But, these corporate officers 
 are not officers of society at large, are not politically elected, are 
 not answerable to the electorate and, therefore, they shouldn't 
 make social decisions. However, they do anyway. 
 
 Student: Certainly a counterargument to Friedman is that 
 corporate independence, the so-called sacredness of the corporate 
 structure, is invalid for all the reasons you mentioned, because 
 the corporation essentially survives within a legal framework that 
 sanctions its activities. 
 
 The option is to either create through their own volition 
 the activities to meet these multifaceted needs or to have them 
 legally imposed. Legal imposition comes, in our society, with 
 actions such as mandating pollution control devices for automo- 
 biles, when the corporate structure isn't responsive to a situation 
 they ultimately affect. The juxtaposition, the interface that's 
 created is much more immediate even if it's more delicate than 
 the Friedmanites would argue. They say that the ultimate bottom 
 line that should make the corporate individual move in a certain 
 way is that it is economically no longer feasible, legally outlawed, 
 or there is a definite societal-legal restraint on a certain type of 
 activity. People are starting to say, "What you have to do is 
 anticipate and respond to problems prior to reaching an impasse, 
 because the sheer inertia of a large corporate empire is damaging 
 to society." 
 
 Dr. Preston: The Friedman people talk about the rules of the 
 game. Neil Chamberlain, who has become a Friedman in sheep's 
 clothing, recently wrote in his book The Limits of Corporate 
 Responsibility: 2 "Oh, yes, you play within the rules of the game, 
 and that means the legal system and what you're authorized to 
 do, but then within the rules of the game you just acquire 
 profits." The question is: What are the rules of the game? How 
 are the rules of the game determined? How are they changed? Do 
 you want the rules of the game to be only a set of legal 
 enactments that are enforceable by the police? That had not been 
 the custom in the free market economy, where the rules of the 
 game have always been much more. Indeed, the cost of a system 
 of game rules which are regulations supported by sanctions is 
 probably greater than society can bear. 
 
 After all, if we have the notion of a private sector structure 
 doing the work of society, accompanied by regulators and 
 controllers from the public sector certifying all the regulations are 
 complied with, we will double the cost of running our society. In 
 that case it would probably be better to put it all in the public 
 sector. It wouldn't be so expensive. 
 
 Besides that, I feel that one of the things we are most eager 
 to preserve in our large diverse society is the possibility of 
 multiple and varied responses. That's a feature of the market 
 system. Anybody can enter the market and offer to do something 
 and if people like it, fine; and if not, it fails. We've got to keep 
 that process going because our society is just too big and too 
 diverse to rely on legislated, regulated uniformity throughout the 
 system. We know from following the experience of other 
 countries, such as the Soviet Union, that it simply doesn't work 
 to run a big complex society in an authoritarian fashion. 
 Authoritarian rules limit the flexibility of response available to 
 deal with problems that arise throughout the system. If you 
 
 2 Neil Chamberlain, The Limits of Corporate Responsibility, Basic 
 Books, New York, 1973. 
 
 172 
 
 accept that there has to be considerable discretion and flexibility, 
 why not use the system that we have, that already includes that 
 discretion and flexibility? 
 
 I think we're going to keep our market system. I don't see 
 any demand for extensive collectivization, and even the tenden- 
 cies that we see toward increased regulation are really matched by 
 counter tendencies. Regulation is so conspicuous that we over- 
 look shifts in the opposite direction, although there are many; the 
 development of alternatives in education is one of them. The 
 whole challenge to the monolithic structure of the educational 
 system is one of the counter tendencies in our society. 
 
 Indicators for Corporate Performance 
 
 There's a tremendous emphasis everywhere on decentraliza- 
 tion and the flexibility it permits. The question that I came 
 to— and this is how I got into the whole subject of social 
 indicators— is, "Suppose the managers of a corporation realize 
 that they are 'socialized.' They know they are in society. They 
 also realize that they cannot become a charitable organization 
 because it would be silly and wouldn't do any good. They are 
 aware that they cannot simply sit there and wait until people 
 scream for this or that and then try to do something. What are 
 they supposed to do?" 
 
 The people that have been thinking and writing about it 
 from a favorable viewpoint have tended to answer that question 
 with lists. They look out at society today and ask, "What is it 
 that people want at the moment?" That is what some of these 
 opinion surveys aim at. They are supposed to provide input to the 
 corporate urban affairs officers and public affairs officers as to 
 what the society around them wants. 
 
 I think that is right, but we need to go beyond that and 
 develop a conception and a methodology that will enable people 
 to handle that question as the answer to what changes society 
 wants over time. If we are looking for a system that will enable 
 the managerial units to continue to function through time, we 
 need a system similar to market forecasting that will not simply 
 produce a list of current goals and desires but, rather, new 
 information about what people will want in the future. In looking 
 for such a methodology that could feed the managerial structure, 
 I offer the social indicators movement. I am now working to 
 develop appropriate reporting analyses that will use the format, 
 methodology, and ideas of the social indicators movement as 
 adapted to the individual managerial unit. 
 
 Let us return to the conceptual picture before we go 
 forward to the data. We visualize the corporate entity, or any 
 other managerial organization, as having a primary function. It's a 
 shoe factory, it's General Motors, or whatever. That is what it 
 does in society. It would not exist if it did not do that. It's 
 nature we call the primary involvement. Around that activity 
 there are all these ancillary involvements that are in the secondary 
 or consequential area. Beyond these secondary involvements is 
 the rest of society. That's a very important point. There are areas 
 of society that you don't touch, or don't touch in any significant 
 way, and these are outside the purview of managerial responsibil- 
 ity. 
 
 If you are going to talk to managers about management 
 they have to be able to discriminate between the items that are 
 on their agenda and items that are not, because they cannot 
 manage the whole world. As a citizen, I can take an interest in 
 anything I want to, but as a manager of an organization, there 
 
SOCIAL INDICATORS FOR CORPORATE MANAGEMENT 
 
 are things that are not on my docket. That doesn't mean that I 
 cannot take some interest in them. 
 
 There are limits of managerial responsibility. Now, the 
 question here is, "How do you organize this? How do you 
 monitor this in order to get some feeling for what society wants 
 you to do and what your effects are; what the goals and standards 
 are in these areas?" 
 
 Basically, there are two kinds of approaches. One is called 
 activity-based scanning, which means that the firm looks at its 
 own penetration into the rest of society, whatever that may be, 
 and says, "I'll follow out what I'm doing from my primary 
 activity. For my primary activity I have to be located somewhere, 
 and that means I'm in a neighborhood. I've got to learn what 
 people want in order to understand my neighborhood impact. 
 Regarding employment, I need to know what people want in such 
 areas as health plans and pensions. I'll trace out my own activities 
 as they touch the rest of society and find out what people say as I 
 
 go." 
 
 The other approach is termed societal overview scanning. 
 Here the firm, instead of starting with its own activities, begins 
 with the rest of society and comes back to the firm. It goes out 
 with opinion surveys and asks, "What is it that society wants? 
 What is the larger society telling me about what units like mine 
 should do?" 
 
 General Electric and AT&T are the two firms that 
 pioneered this. AT&T, which never published the original and 
 very early work that they did more than 10 years ago, has since 
 stopped. Today, GE is the pioneer; it has institutionalized this 
 societal overview scanning and brought it into its very top 
 management levels. In the mid-1960's GE began with a large 
 analysis of trends in modern society, and learned it was not 
 caught up with what was happening. General Electric was lucky 
 with respect to the revolutionary movements of the 1960's. They 
 were not badly knocked about; not nearly as much as General 
 Motors, for example, or many other companies. I think one 
 reason was they had already started in this direction. 
 
 As a successful consumer products company GE must 
 have in its operating procedures methods for understanding what 
 the markets want at any moment in time. 
 
 That fact combined with the electrical conspiracy antitrust 
 case— which was the worst thing that ever happened to General 
 Electric in the public image sense— interacted to bring GE into a 
 very high level of social awareness in the sense of understanding 
 that they exist as an organization within the larger society. 
 Therefore, what society is doing and what it wants and what 
 changes are taking place there are very important. That is just like 
 technological change in an industry. If you are not following that 
 you will not be successful over time. 
 
 Student: With a corporation like General Electric, what you have 
 said seems absolutely unavoidable for its survival because most of 
 the society is its customers. 
 
 Dr. Preston: Don't fool yourself about it being unavoidable. 
 General Motors avoided it for a long time. It avoided it saying, 
 "Mr. Friedman is right," until finally the bombs were going off all 
 around— the Environmental Protection Agency was marching and 
 the Congress was voting against General Motors. Finally it got the 
 message and GM had to start finding out what was going on 
 outside of Detroit. Unfortunately some people still don't realize 
 it. 
 
 Student: To the extent that corporations realize and accept their 
 
 involvement on their own there is no need for government 
 regulation. 
 
 Dr. Preston: Absolutely. Take the General Motors case. When 
 GM realizes it is in the world after all, it can no longer only sell 
 cars, and say that everything is swell and that it is just in business 
 for money. What GM has done, in fact, is activity-based scanning. 
 It has started an inside-the-firm analysis and published the results 
 in a series of annual documents called, "GM's Contributions in 
 Areas of Social Concern." These are quite good. GM holds 
 conferences with its critics and gives them a chance to ask a lot of 
 questions. That's the stage the company is in now and it 
 represents a very sincere and serious response. 
 
 However, the difficulty about activity-based reporting is 
 that it is ad hoc. Even a company like GM only follows out a few 
 of its activities' implications for the rest of society and fails to 
 focus on the total society surrounding it. It doesn't reach any 
 kind of conclusion nor is there any basis for comparison. 
 
 There are other organizations that work in the area of 
 comparisons between companies across industries. The Council of 
 Economic Priorities, in particular, has zeroed in on a few 
 industries and produced some excellent reports. One of the most 
 famous reports, "Paper Profits," concerned the paper industry, 
 and dealt primarily with pollution. After the publicity they did a 
 followup on the paper industry and reported very substantial 
 improvements in a number of companies. They also reported on 
 female employment in the New York banks, a very controversial 
 issue. They are now in court against a corporation to obtain data 
 in order to check on cooperation with the Equal Employment 
 Opportunities Act and related issues. 
 
 While there are people who have entered into this from the 
 critical side, we must also look for ways by which managements 
 can start internalizing on a regular basis. They have to generate 
 some information for their decisionmaking purposes and then, as 
 appropriate or desirable, report to the stockholders, the commu- 
 nity, the complainers, or the employees. After all, these groups 
 are important people and they are interested and affected by 
 these various social involvement activities. 
 
 I was looking for a way to get firms away from the "aren't 
 we good guys" syndrome. That is, companies will report how 
 much they spent on pollution control equipment. But that 
 doesn't tell you a thing about their actual performance in 
 pollution control. Input is a measure of effort, and it's not 
 insignificant, but it's not what we care about. No matter how 
 much was spent, if it doesn't clean up the pollution it was a 
 charade or it was just done to abide by the law, but it didn't 
 accomplish what society wanted. 
 
 The social indicators framework has the virtue of compre- 
 hensiveness. You have a feeling of the whole of society out there 
 and the question is, What aspects of that vast society do you 
 influence? Of course, if you do not influence an area you don't 
 have anything to say about it. 
 
 The social indicators approach is comprehensive. It has this 
 strong emphasis on measurement rather than subjective rhetoric. 
 It also has comparability among measurements. 
 
 If we are going to talk seriously with companies about 
 integrating this social perspective into their management the key 
 is social indicators. It offers a framework that would serve, first, 
 to give the image or the perspective desired. Then, it permits that 
 image to change as society changes. It allows you to get away 
 from the list-of-current-crises approach and into a broader 
 framework. 
 
 173 
 
SOCIAL INDICATORS FOR CORPORATE MANAGEMENT 
 
 Student: How are companies handling the cost of developing 
 these societal information systems? 
 
 Dr. Preston: One of the important things that this approach has 
 going for it is that most of the necessary data are there already. 
 Companies have not brought it together and sometimes they have 
 not even looked at it themselves, but I think we want to 
 emphasize the organization's use of its own existing information 
 in this framework and for comparisons. Some of the indicator's 
 areas have a lot of data, such as employment and job opportuni- 
 ties. Naturally, we have total data; male, female, black, white, a 
 variety of breakdowns under that one indicator. Similarly, on 
 earnings. You have earnings broken down in terms of averages, 
 ranges, and pay scales by a number of appropriate categories. This 
 is one of our most detailed categories. To determine how a 
 company or other organization relates to the rest of society in its 
 income and employment-generating function, look at data for the 
 company and we look at comparable data for its industry, its 
 community, or both. 
 
 One company, for instance, has the interesting problem of 
 an aging work force. That has a lot of long-range managerial 
 implications, many of which are social. For example, there's the 
 question of the medical and pension implications, and the new 
 hiring and training inferences raised by this aging work force. One 
 thing they have come to see better through this process of 
 analysis is that they are going to have to think about some of 
 those implications very soon because a substantial part of their 
 work force is moving into the 60-year age bracket and that's 
 going to have a very important impact on the company. It is 
 going to stretch all through the social involvement areas, because 
 it affects unionization and the turnover rate. I don't think there 
 will necessarily be productivity effects, but there are going to be 
 effects all through the company because of this aging work force. 
 Following that alone through this set of social concerns is a 
 valuable management exercise. 
 
 Student: You mentioned how a corporation, via indicators, 
 could grasp the tenor of its environment both socially and 
 economically. How viable is it for the corporations, and what is 
 the state of the art today, to use publicly available information, 
 Census Bureau data or locally generated data and information? 
 How much could they, and do they, use attitudinal surveys of 
 employees and the local community? 
 
 Data and Reporting Systems 
 
 Dr. Preston: This is an excellent question. It gives me a chance to 
 discuss the report by the First National Bank of Minneapolis. The 
 First National Bank of Minneapolis is not a well-known pioneer in 
 this area. However, they faced exactly this problem. They didn't 
 know anything about the words "social indicators." This project 
 was undertaken by a lawyer who took a job as a vice president of 
 First National Bank of Minneapolis as their public involvement 
 man. He was not content to be just the old fashioned type, "I'll 
 be named to the Board of Directors of the Boy Scouts, and we'll 
 give them $1,000." He asked the very question you asked. "What 
 does the community want, and how can we know things about 
 the community that will be relevant for us?" Without any 
 background at all he set out to conduct a social indicators study 
 for Minneapolis and St. Paul, and he did it. 
 
 174 
 
 The whole story takes 4 years. The first year after he was 
 appointed they announced, in the annual report, the intention 
 to launch into this sort of community survey and listed the 
 areas that they were going to study. The second year, he did 
 what he called the social and environmental audit. Parts of the 
 data were obvious and not hard to obtain, for example, the 
 amount of employment in the city. It was reference point data 
 with some further breakdowns into subgroups. His report had job 
 opportunities, income, health, public safety, housing, education, 
 civic participation, environment (land, air, water), transportation, 
 and culture, 10 areas in all. 
 
 He also wanted to make a Minneapolis/St. Paul and a 
 city/suburbs comparison. In some areas he found that this was 
 almost impossible. In the field of education he wanted perform- 
 ance indicators rather than budget figures. What he really wanted 
 were standardized test scores, results, that are compiled through- 
 out the educational system, which could give readings comparable 
 among schools. Although public record, he simply could not get 
 that and had to give up. He went to drop-out rates, as they are 
 generally thought to be of some importance for social cohesion, 
 and they reflect job opportunities. Also, it is indicative of the 
 effectiveness of the educational system, since if the system were 
 really effective people would stay in it until they were finished. If 
 students are dropping out it is a sign that something is wrong. He 
 almost had to go to court to get the data on the drop-outs from 
 the school districts. Sometimes it's really less a matter of 
 availability in the sense of the existence of information than 
 releasability of the information from the people who have it. 
 
 He was very fortunate to be in Minneapolis-St. Paul because 
 that is a fairly sophisticated area. It is an educated, well-managed 
 community and they had already witnessed several studies. One 
 was a poll on job attitudes, showing the percent of people 
 dissatisfied with their jobs. Other polls reported people who 
 thought they had access to adequate health care; who felt unsafe 
 walking the streets at night; who were willing to have low 
 income housing in their neighborhoods; who believed the 
 suburban schools were superior to the city schools; and data on 
 participation in the United Fund. He was able to gain access to 
 several studies of an attitudinal nature. From this compilation of 
 data and without any new data collection activity whatsoever he 
 produced a very good community profile. 
 
 The next year, 1973, he took these same topics, but not the 
 same indicators, connected them to the bank and asked how the 
 bank's performance was related to these areas. For example, in 
 the housing area, his questions had to do with the loan program 
 of the bank. That was the consequential involvement, the way 
 that the bank was connected with housing, not its own housing, 
 but its impact on community housing. Certain items concerned 
 employees, such as their wages. Involvement was analyzed, as 
 with the employees' contribution to the United Fund, which was 
 taken as a sign of community cohesion. They had some survey 
 data about job satisfaction for the bank and for the community. 
 What he did next was to assemble these available data from 
 inside. Again, as far as I know, no new study of anything inside 
 the company was made. It was just a question of pulling out the 
 things that were there and drawing them into this framework. He 
 developed a priority ranking of issues for the community on the 
 basis of a city/suburbs disparity. He adopted the principle that 
 the elimination of this disparity would be desirable. Areas where 
 there was great disparity between the city and the suburbs 
 indicated needed work, and he used that as his priority ranking. 
 By using the data he tried to indicate areas where he thought the 
 
SOCIAL INDICATORS FOR CORPORATE MANAGEMENT 
 
 bank was making an optimum contribution, just maintaining the 
 status quo, or having a detracting impact. He did not advise the 
 bank how to answer the questions the report raised. What he 
 wants to do now is what I am doing, which is to match the 
 indicators so that you are comparing those for the bank directly 
 with the ones for the community. 
 
 In our own work, we have been pleased with our progress. 
 We have about 50 indicators, though they aren't all matched yet. 
 We have some for the company, others only for the industrial 
 community. We are in pretty good shape because we are using 
 both the community and industry to develop a basis for 
 comparison. For these terms we use rather flexible definitions. 
 We may use industry group, which is a very general definition, if 
 that is the best thing we can get. Or we may use a specific 
 industry. We may even set up a two-firm comparison, if 
 appropriate. With the community, sometimes we use U.S. data, if 
 available and relevant, for life expectancy or similar items. In the 
 product-quality consumer-concern area we employ information 
 about law suits or consumer complaints. There may be data like 
 that on some type of national basis or we may use the State, 
 county, or city, depending on what appears relevant. The whole 
 process is designed to enable the management of the company to 
 see itself in this social context and to observe how its impact 
 compares with what is going on in those areas where we believe 
 society is interested in the results. 
 
 Let me give you one more example of what we find. I was 
 determined to work with a manufacturing rather than a financial 
 company because financial companies are too easy and their 
 impact is too indirect. My example is a manufacturing company 
 with real safety problems. Its manufacturing procedures are such 
 that there are some tasks that are very difficult to make clean and 
 safe. They have done a lot to improve the safety situation, but it 
 is not by any means perfect nor is it a perfect environmental 
 situation. We started talking about accident rates with the 
 management and they responded that such information was very 
 confidential. Everything we are doing for them is confidential. 
 We're not releasing it to anyone without their approval, and we 
 consider ourselves inside instead of outside. They have to report 
 accidents to their insurance company and to the State. How 
 confidential is all that? Also, as private citizens we could go to 
 the State and sue, if necessary, to obtain some of this data. We 
 finally got them to stop pretending about confidentiality. It 
 turned out that they thought that their record might not be 
 favorable. 
 
 Eventually, of course, we got the data and made compari- 
 sons with all manufacturing units in the same industry. Our 
 company has an outstanding safety record. Perfect safety, of 
 course, means no accidents. They don't have that. We don't know 
 anybody that does. But in comparison to things that are relevant, 
 their safety record is something they can brag about. There 
 certainly is no reason to suppress the safety record. Yet, they 
 didn't even know that, because they hadn't put their safety in 
 perspective. I think that as they become more familiar with this 
 way of looking at things they may overcome the feeling that 
 people are going to find out something bad about them. Of 
 course, sometimes you do find out something bad. We found out 
 that one of their plants is under EPA sanctions for air pollution 
 discharges, although they say they are working very hard to clean 
 it up. I'm sure they are working hard to clean it up, but the fact 
 of the matter is that out of "x" number of plants one is 
 sanctioned while the rest are not. I feel that in the process of 
 doing this kind of work you cause people to become more 
 
 accustomed to acknowledging the whole truth. This is the same as 
 with financial records where you don't simply report the sales for 
 years they are good, but you report them every year. 
 
 As we move toward institutionalizing this type of approach, 
 we'll overcome the feeling that characterized financial accounting 
 100 years ago. "No, we don't reveal our sales to anyone." You 
 just couldn't run a modern business that way. We are at the stage 
 where many things about the operating characteristics of a firm 
 and its impacts on society are just not going to be secret 
 anymore, because they're not really secret anyway. They're all 
 filed somewhere and the facts are available if you go and get 
 them. But the management is just coming to face this. 
 
 We have been hard-pressed to find comparison norms all the 
 way along the line. Indeed, it is comparison norms that today are 
 our biggest problem. I thought, when we started this project, that 
 we could pretty well turn to public sources, particularly since 
 Buffalo was covered by the Urban Institute Study. I thought we 
 had a reasonably good profile on Buffalo right there. Plus, 
 knowing how much census data are available on standard 
 metropolitan statistical areas and central cities, I believed it 
 would really be quite easy to generate at least the community 
 comparisons. We have found that for this particular purpose- 
 because we do want to have the indicators that match something 
 for a micro-organization— we must attempt considerable refine- 
 ment and we are working on a social indicators project for the 
 Erie-Niagara SMSA which is specifically aimed at this use. 
 Hopefully, we can encourage other companies to use it for their 
 own analysis and to generate their own internal data. 
 
 I think that the application of these techniques to the 
 university is very important, where the university itself is a large 
 entity, because it has impacts on so many of these areas, like 
 housing, for example. It has so many environmental impacts. It 
 has employee impacts. All of these things are consequential to its 
 primary educational-research function. This whole approach is 
 applicable to any other kind of large multifunction organization. 
 This whole approach may not be applicable to a hospital, let's 
 say, or a health facility, unless it is very, very large in its 
 community. That kind of unit can more appropriately focus on 
 one or two areas that would be relevant for its operation and lots 
 of other areas could justifiably be left blank. 
 
 The problem that has plagued even the most sincere and 
 precise attempts at corporate social analysis is that companies are 
 all different in the first place, they don't generate ways of looking 
 across a group of companies. They're ad hoc. You do it this year 
 and you don't do it next year. 
 
 I feel that the corporate sector, in particular, is looking for 
 a technique, and it has got to be connected with a theory, with an 
 image, with a concept of what a corporation is. They're looking 
 for a technique that will embody a concept that they can adopt, 
 apply, and change over time. There is a real need for this kind of 
 approach and I think that the social indicators theme is a viable 
 one for this purpose. Hence, I am interested in the development 
 of the indicators and particularly in their development on the 
 local area basis, because that is where a corporation operates. By 
 and large, it doesn't have an impact on the whole society. They 
 have most of their impact where they are located. 
 
 Student: Has the Minneapolis bank taken action on the vice 
 president's study? How does he intend to use his information and 
 are the uses of the information somehow implicit in the goals he 
 established? 
 
 175 
 
SOCIAL INDICATORS FOR CORPORATE MANAGEMENT 
 
 Dr. Preston: He says that this system they have developed has 
 affected a few of their major policies, as in the housing mortgage 
 area and in connection with transportation planning in Minnea- 
 polis. That is, it has affected what the bank as an entity would 
 do, what stand it would take about changing the transportation 
 system and things of that sort. 
 
 Student: It seems the answer to my question is that the goals 
 which determine the selection of the indicators imply the 
 ultimate usage. As an organizer, you have to know in advance 
 what you want, why you want it, and where you feel your 
 responsibilities lie. 
 
 Dr. Preston: No, I don't think you have to know the answer, but 
 you have to know the character of the answer. The character of 
 the answer is that you want to manage this organization in a way 
 that is viable and hopefully as close to optimal as you can get in 
 relation to the larger society. That includes producing goods 
 economically and selling at a profit, too. I think that's amoral, it's 
 neutral. But, I think it's a managerial error to exercise in the role 
 of a manager one's own moral judgment when that differs from 
 the larger society. 
 
 Student: What about maximizing certain values, don't you have 
 to know what values you're after? 
 
 Dr. Preston: I really don't think that anybody operating in any 
 complex multifunctioned activity maximizes anything. We're 
 back to the image. I think maximization is the wrong image for 
 management behavior. It's very important, within management, 
 to see what it would take to maximize one thing and minimize 
 another. However, I think that in conducting the total operations, 
 whatever they are, you never maximize anything. There are mul- 
 tiple goals; there always are, there always have been; and what 
 
 you do is to adopt a strategy that involves a combination of these 
 goals, no one of which is ever maximized. 
 
 Student: Let's say health or quality of life? 
 
 Dr. Preston: You'll never maximize it. You can't maximize the 
 quality of life because that has too many dimensions, so it's just 
 not possible. The only way to maximize health is to climb into 
 the intensive care unit and get all hooked up so that you'll be 
 kept in perfect health at all times. That's like shooting the 
 unemployed. You don't want to maximize health. You certainly 
 would never go out on the street, you might get hit with 
 something! And in Los Angeles, even breathing, the air is 
 dangerous. 
 
 I think that the impact on management thinking of the 
 economic notion of maximization has been bad. On management 
 practice it has been good, but on management thinking it has 
 been bad. It keeps people from understanding what they are, in 
 fact, doing. They are not, in fact, maximizing. It's a mistake to 
 think of maximizing things when you're not. 
 
 Student: What you're really saying is that it's better to intelli- 
 gently allocate your resources than abdicate that allocation. You 
 can make good or bad judgments, but you should make 
 intelligent judgments, whether they're good or bad. 
 
 Dr. Preston: You try to do the best you can. But to move toward 
 the kinds of judgments that will be successful you need to 
 understand the organization as an element in society, an 
 organization in the social context. It's not a question of whether 
 they'll be morally good or not. It's a question of whether they 
 will be successful. Successful means not having riots, not having 
 your plant burned down; but it also means having people buy 
 your product and your people come to work. If an organization 
 does not have those characteristics it is not successful. 
 
 176 
 
The Choice of Methodology In Social Problem Solving 
 
 Harold Nisselson, Chief Mathematical and Statistical Advisor, 
 U.S. Bureau of the Census 
 
 "I don't mean to say we need a model of the whole world before we can do 
 anything; but every once in awhile you have to lift yourself above the trees and see how 
 your measure of effectiveness is related to the total objectives by asking what you are 
 really trying to accomplish with the program." 
 
 Harold Nisselson 
 
 Mr. Nisselson: To set some context and scope to this morning's 
 seminar I have taken the approach that the people in the graduate 
 school of public administration are present or prospective city 
 managers, or people with the responsibility for operations and 
 problem solving. My purpose is to talk about the general design of 
 analytic systems for social and economic problem solving. 
 
 I want to cover quantitative and other methods which are 
 objective and provide measures of uncertainty, or at least 
 measures of sensitivity, to the assumptions made within and the 
 results of the analysis. I exclude subjective methodologies and, 
 particularly, argument by counterintuition, as well as any 
 problems of national accounts or models of the economy. 
 
 The problem that typically faces decisionmakers in the 
 context of social and economic problem solving is the allocation 
 of resources between competing demands or alternative invest- 
 ment opportunities. Decisionmakers need to know how much of 
 what kind of benefits or outputs can be bought for a given target 
 group and at what cost. I believe that systems analysis is a useful 
 tool for this purpose, because it provides estimates for the 
 alternatives and a basis for analysis of cost-effectiveness. The 
 classical paradigm of hypothesis testing is less applicable to the 
 problems of decisionmakers because it does not. Given that a 
 statistical test shows one alternative to be better than another, 
 the question still remains: How much better, and at what cost? 
 In my opinion, the questions really pose problems of estimation, 
 not tests of statistical significance. 
 
 First, a system is a set of interrelated elements and some 
 transfer functions which relate the inputs of elements to their 
 outputs. These elements are all related in trying to accomplish 
 some objective of the system. For analysis purposes we need to 
 know how much input to an element leads to how much output, 
 in going from one element to another and at the end. That, of 
 course, is the ideal situation. Frequently we will fail to achieve 
 the ideal, and quite often the best we can do is approximate it. 
 
 What we want to do is to talk about the general paradigm 
 that allows the recognition or identification of a problem in 
 relation to an objective. The term "system analysis" encompasses 
 this approach. Somehow we become aware of the fact that a 
 problem exists. Awareness is achieved by events ranging from the 
 burning down of a city, or complaints to an agency, to just 
 becoming aware of the fact that some work isn't getting done. 
 
 To carry out a system analysis we need a specification of 
 the system's objectives, requirements, or desired outputs. We need 
 to know what constraints may affect the system or solutions. 
 Some solutions may not be politically acceptable, some may not 
 be feasible for one reason or another, while others may be cost 
 prohibitive. 
 
 With an awareness of these points we design some alterna- 
 
 tive solutions to achieve the desired output or end result and 
 establish criteria for evaluating the alternatives. 
 
 Then, there is experimentation with the alternatives and 
 analysis of the results. One alternative, typically, will be to 
 maintain the present system, since it is always possible that a new 
 system will make things worse. 
 
 As I have said before, the analysis is intended to lead to the 
 ideal solution. Because the ideal can't be achieved, the most 
 practical solution must be chosen. 
 
 Student: Could you give an example of a transfer function? 
 
 Mr. Nisselson: The simple type that many of us deal with is how 
 much lawn grass is grown for how much fertilizer applied. 
 There are relationships that tell you how much fertilizer to use 
 per square foot, and you have a model for the lawn, length times 
 the width to get the area. But, because the lawn is really rough, 
 the area is only an approximation. Generally we understand these 
 transfer functions in the physical sciences a lot better than those 
 in the social sciences. For example, we don't know how much 
 training input to unemployed or unemployable persons will result 
 in how much employment. 
 
 The analysis compares the alternative solutions with regard 
 to the measure of effectiveness. Frequently, the case is that there 
 are multiple objectives, and then trade-off analysis is required to 
 compare the compromises that are attainable with regard to each 
 objective. 
 
 The following example illustrates in a more formal way the 
 kind of analysis that we do intuitively. This one deals with refuse 
 collection in a city of about 100,000 population. 
 
 The refuse is picked up weekly on a regular schedule. The 
 refuse crew consists of a driver and additional men who pick up 
 the trash and set it on the street, someone dumps the trash into 
 the truck, and one or two men replace the trash cans. When the 
 truck is full, it drives to the incinerator and returns. When they 
 finish the day's assignment the truck makes a last trip to the 
 incinerator, and the other people go home. 
 
 Some of the questions for analysis are: What kind of truck 
 is required? Should you consider mainly the matter of the size or 
 the capacity of the truck? What should be the size of the crew, 
 how many people should you have on the different operations, 
 how many people altogether? What is the design of the route, 
 what kind of shape should it have, what kind of coverage, what 
 kind of mix or kinds of houses? How should the routes be 
 scheduled? What is the maintenance and replacement policy for 
 equipment, which is essentially when should you replace trucks? 
 Then, you could consider system alternatives. 
 
 177 
 
THE CHOICE OF METHODOLOGY IN SOCIAL PROBLEM SOLVING 
 
 The analysis that's illustrated here deals with the truck 
 replacement policy, and it's very much like the analysis that you 
 go through when you consider trading in a car. On the one hand, 
 the longer you keep a car the lower the depreciation is, and so 
 you're amortizing your initial investment. On the other hand, the 
 longer you keep a car the more repairs you have to make on it. 
 
 Figure 1. ANNUAL OPERATING COST VS. AGE FOR 
 REFUSE TRUCKS 
 
 Figure 2. PROBABILITY OF BREAKDOWN VS.AGE 
 FOR REFUSE TRUCKS 
 
 o 
 
 o 
 ©_ 
 
 T— 
 
 o 
 O 
 
 Age, years 
 
 Repair costs are reflected in figure 1 which shows a rough 
 plot of the annual operating cost against the age of this type of 
 truck. This increasing cost is a reflection of the fact shown on 
 figure 2 that the probability of a breakdown keeps increasing 
 with age. Finally, if you plot the days of use you get out of a 
 truck per year compared to the age, you see that the usage of a 
 truck goes down. There was some consideration of what the 
 measure of effectiveness should be, and the one chosen was not 
 annual cost but cost per day of use. That's plotted in figure 3. 
 This shows a typical shape, as the cost per day of use declines to a 
 minimum and then goes up again. 
 
 Basically, there's only one way to evaluate a program design 
 and that is by running some controlled experiments to compare it 
 with the alternatives or with the absence of a program. These 
 experiments can be very expensive and time consuming, so it is 
 exceedingly useful to be able to approximate the results of 
 experiments with paper and pencil analysis. And in cases where 
 the paper and pencil analysis isn't a complete representation or 
 doesn't appear to be the final answer, and you still want to have a 
 real life test, the analytic results can be used to help guide the 
 type and design of the field experiment to make the process more 
 efficient. For example, the results could be used to eliminate 
 designs not even worth testing. 
 
 178 
 
 
 
 
 
 
 
 — i 
 
 c cu 
 
 .25 
 
 
 
 
 
 /< 
 
 o -* 
 
 
 
 
 
 • / 
 
 
 TJ c 
 
 
 
 
 
 
 
 reak 
 ay i 
 
 .20 
 
 
 
 
 
 
 CO Q 
 
 
 
 
 
 
 
 °i 
 
 .15 
 
 
 
 
 
 
 >• o 
 
 .r -o 
 
 
 
 
 
 
 
 — -* 
 
 
 
 
 
 
 
 CO <u 
 
 .10 
 
 
 
 
 
 
 Prob 
 in Br 
 
 .05 
 
 
 
 
 • 
 1 1 
 
 
 1 
 
 
 1 2 3 4 5 6 
 
 Age, years 
 
 Figure 3. COST PER DAY OF USE VS.AGE FOR 
 REFUSE TRUCKS 
 
 To carry out this sort of analysis we need a mathematical 
 model, a set of equations or relationships and factors which 
 abstracts the essential features of the real world system to a 
 satisfactory degree of approximation. 
 
 In simple cases, an explicit mathematical expression for the 
 solution can be obtained. However, when the analysis becomes 
 too complex, a computer is employed to project all the 
 possibilities and record the results. This is referred to as a 
 simulation model. 
 
 Some idea of the power of such a method is shown by the 
 observation that several years of operation of the health care 
 
THE CHOICE OF METHODOLOGY IN SOCIAL PROBLEM SOLVING 
 
 clinic we will talk about could be simulated in about 4 to 10 
 minutes of computer time, depending on the type of computer. 
 And, this is a powerful savings. 
 
 I might say that the design of simulation models is a 
 speciality all of its own. You may have run across references to it 
 in war games, analysis of big weapon systems, and tactics. A lot 
 of the developments came from military applications. 
 
 Mr. Barabba: A valid criticism about models is that on the one 
 hand they are used to manipulate and analyze data more 
 efficiently and effectively and with less bias than a person could. 
 But, a point can be reached where the data is bad or the theory 
 behind the data selection is not understood, and rather than 
 stopping there, the model builder works around the bad data and 
 goes on to the end result. When this is done there is a need for 
 criticism. 
 
 What model builders should do when they reach an obscure 
 point is to stop and say, "This is how far we've come. Here's all 
 the information available. Now, let's have the decisionmaker 
 select the next data element on the basis of experience, 
 knowledge, and his understanding of the problem." The decision- 
 maker may be wrong in his selection, but at least he is aware of 
 exactly what input resolved the difficulty. 
 
 Mr. Nisselson: That's a very, very important observation. I would 
 be very suspicious of someone who couldn't explain the logic of 
 what he was doing. 
 
 Mr. Barabba: Models cause us to think systematically. They're 
 really an asset. I recall that we were asked to design a political 
 voter behavior model, and to do this we brought in political 
 campaign managers from all over the country but not one 
 political scientist. We asked them, "What do you look for when 
 you have to make a decision?" What kind of decisions do you 
 have to make?" It ended up that most of the items needed to 
 make a good political decision were available, either from survey 
 research or records of past voting behavior. 
 
 And for the other items where you didn't have information 
 we asked, "What's the concensus of the group?" The people for 
 whom we designed the model were very comfortable with this 
 procedure as they knew the group from which the concensus was 
 drawn, and that group had either won elections or been involved 
 in good decisionmaking. The model we developed ended up as a 
 very, very accurate predictor of activity. 
 
 Even more important is to say, "What if?" That's really the 
 key here because that's when you really find yourself with a 
 valuable piece of information. You're going into a situation and 
 you have a couple of alternatives and you say, "What if it was 
 this? What if it was that?" Once you have the answer to that 
 question with a degree of confidence around it, you've really 
 made a major contribution to a decision. The systematic 
 description of the phenomenon that takes place is a requirement 
 of a good model. In other words, it forces you to plan and to 
 think systematically. I think some of the better models don't go 
 through the computers. 
 
 Student: It appears to me that the model designers might be 
 prejudiced to what data are available in order to make their 
 model work. 
 
 Mr. Nisselson: That's a troublesome question. There's no sense 
 developing a model for which you can't find data to use. 
 
 Sometimes people fall back on what they call sensitivity analysis. 
 They say, well, I'll invent some data and see which are the most 
 sensitive factors. On the other hand, one advantage of a good 
 model, if you really understand the system and have the right 
 variables, is that it tells you what information to collect. If you 
 just go around scooping up data that are around you may get 
 variables that aren't very meaningful or very useful. In a systems 
 analysis study you will frequently find yourself going out to 
 collect special data that you need. 
 
 A crucial point in comprehensive systems analysis is the 
 choice of the measure of effectiveness. 
 
 For example, we could analyze a library system and its 
 efficient operation. But if we fail to evaluate it in terms of the life 
 or the social problems of the community and the community's 
 objectives, we fail to have a systems view of a library in the best 
 sense. 
 
 That was the observation that Professor Preston was making 
 yesterday when he was objecting to maximization. He was saying 
 that the measures of effectiveness, if they're badly chosen, merely 
 talk about the efficiency of a single system and don't have 
 anything to do with what you're trying to do in the community. 
 If you can somehow find those connections between the system 
 and the community, however you define "community," and 
 relate them to the kind of community you're trying to build, or 
 the problems that you're trying to solve, then you'll have a much, 
 much more effective kind of analysis. 
 
 For example, in a probation or parole system, you can 
 make the system look very good by just not allowing any risky 
 cases to be paroled, and thus there is less recidivism; you never 
 have people breaking parole. Except, that isn't necessarily what 
 you want to do. 
 
 In a manpower training system a commonly accepted 
 measure of the effectiveness of the system is the number of 
 people placed in jobs. Well, the fact of the matter is that most 
 manpower programs deal with very few people in the labor force. 
 The programs have been criticized for placing participants and 
 displacing workers not enrolled in these programs. Unless you've 
 done something to create more job opportunities, maybe all 
 you've done is just push the vacancies around, this measure that 
 you're so proud of doesn't really measure anything. 
 
 Mr. Barabba: However, it is important that you not confuse the 
 measure of effectiveness. For example, a congressman critically 
 points out all the defaulted FHA government-subsidized mort 
 gages. That's to be expected, because that is part of what the 
 program is supposed to do. It's supposed to take credit risks and 
 give them an opportunity to have a home. If there were no FHA 
 mortgage defaults that would mean the program was probably 
 dealing with the people who are not credit risks, rather than those 
 it was created to assist. If it deals with the safe market, it has 
 taken over a segment of the market place that the private sector 
 could have dealt with. 
 
 Mr. Nisselson: It's particularly important to continually monitor 
 the system because people within the system often pervert the 
 measures of effectiveness. 
 
 Analysis of police operation are like that. If you measure 
 efficiency by how many arrests are made, they'll make arrests. If 
 you measure convictions, a way is found around that. Things are 
 made to look less serious; grand larceny is downgraded to petty 
 larceny. Or, people are persuaded not to file complaints. It is 
 necessary to run quality control checks on the records to make 
 sure the measures and systems are not being perverted. 
 
 179 
 
THE CHOICE OF METHODOLOGY IN SOCIAL PROBLEM SOLVING 
 
 I don't mean to say we need a model of the whole world 
 before we can do anything; but every once in awhile you have to 
 lift yourself above the trees and see how your measure of 
 effectiveness is related to the total objectives by asking what you 
 are really trying to accomplish with the program. 
 
 This next example illustrates a different kind of a point. 
 The analysis was developed in connection with a North Carolina 
 comprehensive maternal infant health care program centered in 
 three counties. Presumably, however, with some appropriate 
 modifications this method would be generally useful as a tool for 
 analyzing maternal and infant care programs anywhere. 
 
 In this case, the authors were very conscious of the fact 
 that there was a need to test the results of the model with some 
 actual experience. The model was developed over 2 years and 
 tested by comparing what the model said was going to happen 
 with the actual results for the following year. 
 
 The analysis report states, "In general, the model produces 
 an acceptable fit for the Wayne and Halifax County operations. 
 The Warren County data does not indicate a good fit." Now they 
 began to look at why this was. And, when things don't work, 
 looking at why they don't work is a very informative process to 
 go through. In discussions, the project personnel said, "Well, 
 somehow it's different from the others." There was a different 
 feeling about the operation, a different attitude towards the 
 clinic, maybe people took it more into their lives and used it 
 
 more. Their conclusion was that there was an extra variable which 
 was not included in the model, and if you wanted to improve it 
 you'd have to do further studies. The fact of the matter is that 
 models like this have been further developed and explored, and 
 they have been used in planning maternal and infant care clinics. 
 
 All of this ties in with the Census Bureau. In analysis there 
 are all kinds of numbers that you need. You say, "Where do they 
 come from?" They come from basic population data, they come 
 from the Census Bureau. Some may come from special surveys, 
 while some of them are generated by administrative records for 
 operations of the county system. 
 
 You could get things like fetal deaths, for example, out of 
 the vital statistics system. You get the denominator, that is, the 
 number of births of which there was a risk of fetal death by the 
 records of deliveries made in hospitals. Fortunately, in that case, 
 we find that a very high percentage of births are in hospitals— 
 about 99 percent of registered births take place in hospitals— so 
 that provides a good base. 
 
 One thing you find is that the great need in all of these 
 things and especially in the case of social indicators is to find 
 appropriate denominators for the numerators that are generated 
 by administrative record systems. 
 
 Reverting to the example, this analysis is intended to 
 represent an idealized view of the real world as shown in figure 4 
 
 Figure 4. THE GENERALIZED MIC PROJECT HEALTH SERVICE SYSTEM 
 
 Community Population 
 
 Associated Attribute 
 
 Generation 
 
 Potential Mothers, 
 Children 
 
 Health Needs 
 
 • Pregnancy Generation 
 
 • Infant Illness Generation 
 
 INFANT CARE CLINIC 
 
 N. C. STATE BOARD 
 OF HEALTH 
 
 Health Resource Planning 
 • Allocation of Resources 
 to Various Subsystems 
 
 Pregnancies 
 
 Health Need to 
 Demand Conversion 
 
 Case Detection by 
 Public Health Nurses, 
 Social Workers 
 Clinic Information 
 Spread by Informal 
 Leaders, etc. 
 
 PROMOTION 
 
 
 
 INFANT CARE CLINIC 
 
 
 
 
 
 
 • Diagnosis 
 
 • Treatment 
 
 • Prevention 
 
 • Promotion 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 1 
 
 ' 
 
 
 
 
 
 
 PEDIATRICIAN OFFICE 
 
 
 
 • Diagnosis 
 
 • Treatment 
 
 • Prevention 
 
 • Promotion 
 
 HOME 
 
 
 
 c - 
 
 • Rehabiliation 
 
 
 
 
 
 - 
 
 1 
 
 
 
 - 
 
 \ 
 
 
 
 
 
 
 
 
 
 HOSPITAL OUT-PATIENT 
 CLINIC 
 
 
 
 
 
 HOSPITAL 
 
 
 
 • Diagnosis 
 
 • Treatment 
 
 • Prevention 
 
 • Promotion 
 
 
 
 
 
 • Rehabiliation 
 
 
 
 
 
 
 
 
 
 
 
 • Not Full Term 
 
 180 
 
 
 MATERNAL CARE CLINIC 
 
 
 
 • Diagnosis 
 
 • Treatment 
 
 • Prevention 
 
 • Promotion 
 
 
 
 
 
 
 
 
 
 
 
 HOSPITAL/HOME 
 
 • Delivery and Care 
 of Mother 
 
 • Care of Child 
 
 Potential 
 
 Future 
 
 Pregnancy 
 
 FAMILY PLANNING 
 CLINIC 
 
 • Promotion of 
 Family Planning 
 
THE CHOICE OF METHODOLOGY IIM SOCIAL PROBLEM SOLVING 
 
 Figure 5. INTERRELATIONSHIPS OF MODEL ROUTINES 
 
 START 
 
 SIMULATION 
 
 (START) 
 
 CHANGE 
 
 CERTAIN 
 
 PARAMETERS 
 
 (CHANGE) 
 
 END 
 
 SIMULATION 
 
 (ENDSIM) 
 
 REPORT YEARLY 
 OPERATION 
 STATISTICS 
 
 (REPORT RSUMMA) 
 
 GENERATE 
 
 PREGNANCIES 
 
 (PREG) 
 
 SCHEDULE MATERNAL 
 
 CARE AND ALLOCATE 
 
 CLINIC RESOURCES 
 
 (MCARE) 
 
 REPORT MONTHLY 
 MATERNAL CLINIC 
 
 STATISTICS 
 (REPORT RMCARE) 
 
 SYSTEM 
 MONITOR 
 
 REPORT YEARLY 
 
 SUMMARY OF 
 
 PREGNANCY 
 
 TERMINATIONS 
 
 (REPORT RPREG) 
 
 SIMULATE 
 BIRTHS 
 (BIRTH) 
 
 SCHEDULE INFANT 
 
 CARE AND ALLOCATE 
 
 CLINIC RESOURCES 
 
 (ICARE) 
 
 REPORT MONTHLY 
 
 INFANT CLINIC 
 
 STATISTICS 
 
 (REPORT RCHILD) 
 
 It's pretty clear how you start out through the boxes. There's a 
 community population in the upper left-hand corner and there 
 are some associated attributes generated. 
 
 If you take mothers by marital status, age, sex, race, etc., 
 you have pregnancy rates formed so you can generate events (i.e., 
 pregnancies) and determine whether the pregnant woman uses the 
 clinic. Given all those circumstances you ha^e factors or 
 equations that tell you how much stillbirth you'll have, how 
 much fetal death, and after a child is born, how much infant 
 illness. You can go through all these boxes from one to the other 
 and, cycling through, you can see how the analysis works. The 
 analysis logic, the way we're going to analyze this abstraction, is 
 shown in figure 5. A summary of the model operations is given in 
 table 1. 
 
 A system monitor keeps track of everything that's happen- 
 ing. You start the simulation which gets the monitor going, and 
 generate pregnancies. Then you can schedule maternal care and 
 allocate the clinical resources. Births are simulated, infant care 
 schedules and so on. You go through all of the steps in the 
 analysis and keep track of what happens. 
 
 The "four measures of effectiveness considered are given as 
 follows: 
 
 1) The maternal clinic attendance ratio: the number of 
 scheduled visits divided by the number of actual 
 visits. Normally an attendance ratio is expressed as 
 "actual divided by scheduled"; however, in order to 
 
 make all effectiveness measures indicate improve- 
 ment as they become smaller in size, this inverse of a 
 normal ratio was adopted. 
 
 2) Total personnel utilization ratio: the ratio of total 
 hours utilized divided by hours allocated. When 
 greater than one, this ratio indicates overutilization 
 and, thereby, poor attitudes on the part of clinic 
 patients (since they are receiving less attention than 
 when the ratio is less than one). However, as the ratio 
 becomes smaller than one, it indicates an increased 
 cost per patient treated. 
 
 3) Abnormality ratio: the ratio of project abnormalities 
 divided by project population, to population abnor- 
 mality divided by the total population. Population 
 abnormality was based on the area as a whole. 
 
 4) Infant illness ratio: the number of project infant 
 illnesses as a percent of total project infant patients 
 divided by the equivalent population illness ratio for 
 the area as a whole. 
 
 Two of them dealt with clinic operations and the last two 
 dealt with health outcomes; one of them being birth abnormal- 
 ities which are miscarriages or fetal deaths and the other one, 
 infant illnesses. 
 
 Typical results, outputs, that this analysis would provide 
 are shown in the tables in figure 6, which summarizes the first 
 
 181 
 
THE CHOICE OF METHODOLOGY IN SOCIAL PROBLEM SOLVING 
 
 three effectiveness measures for a year's operation of the 
 program, and figure 7, which summarizes the statistics on the 
 outcomes of pregnancies. Other outputs from the simulation 
 show details of utilization and costs of clinic staff by type- 
 physician, nurse, social worker, and so on. An example of a 
 
 comparison is shown in figure 7 where you see that, for the 
 community as a whole, miscarriages and fetal deaths are running 
 about 9.2 percent and for program patients they were running 
 about 3.3 percent. The program, as it was run according to those 
 parameters, was a great success. 
 
 Table 1. DESCRIPTION OF MODEL OPERATIONS 
 
 MODEL IMPUTS 
 
 Population Class Characteristics 
 
 including size and crude pregnancy 
 rate. 
 
 Characteristics of the Pregnancy 
 
 including, for each class, age dis- 
 tribution for mother, marital 
 status distribution, parity distri- 
 bution, and pregnancy outcome 
 distribution. 
 
 Initial Visit Distribution 
 
 indicating during which month of 
 pregnancy the mother attends the 
 clinic for the first time. 
 
 Economic and Educational Distribution 
 for mothers attending the clinic. 
 
 Resource List 
 
 The personnel resources used in the 
 maternal and infant care clinics in 
 addition to the pediatrician office 
 resources . 
 
 Hospital-Stay Distribution 
 
 for both mother and child at the 
 time of delivery. 
 
 Cost Data 
 
 for the clinic and hospital opera- 
 tion 
 
 Infant Illness Data 
 
 based on project experience includ- 
 ing illness outcome, where treated, 
 and conditional probability of hav- 
 ing subsequent illness. 
 
 SIMULATION OF THE MATERNAL AND 
 INFANT CARE PROGRAM 
 
 The Model: 
 
 Calculates the number of pregnancies 
 generated in a particular month and 
 assigns characteristics to them. 
 
 Builds a schedule of clinic visits for 
 the mother. 
 
 Allocates resources for these visits 
 and 
 
 Records their utilization and costs. 
 
 Creates a record for each pregnancy 
 that terminates in a live birth and 
 assigns appropriate characteristics 
 to this infant's record. 
 
 Generates an illness pattern for the 
 infant and 
 
 Builds a schedule of both clinic and 
 office visits by the mother and in- 
 fant. 
 
 Allocates resources for these visits 
 and 
 
 Records resource utilization and costs. 
 
 MODEL OUTPUTS 
 
 Maternal Clinic Utilization Report 
 
 for each month which indicates case 
 load, amount of resources used, 
 maximum clinic running time 1 , and 
 cost of clinic operation. 
 
 Infant Clinic Utilization Report 
 which is similar to the Maternal 
 Clinic Utilization report. Data 
 for the infant clinic, the hospital 
 and pediatrician office visits are 
 reported. 
 
 Yearly Summary Report 
 
 summarizes the clinic operations 
 for the year and includes effec- 
 tiveness measures. 
 
 Pregnancy Termination Report 
 
 a yearly report which indicates the 
 number of pregnancies that termi- 
 nate during the year and their out- 
 come, both for the target popula- 
 tion and for the population at 
 large. 
 
 Student: Is this the way you would present it to the adminis- 
 trators, the decisionmakers? 
 
 Mr. Nisselson: You would actually go through some trial and 
 error showing them different ways of looking at it. You'd 
 probably have had enough discussion and have some know-how 
 before hand. You'd know something about what people were 
 looking at and what they ought to look at. It's a continuing 
 process of finding out how to communicate with your boss. How 
 do you present things to your boss in a way that they'll hit him 
 right? If you get a new boss, you have to find out how he likes 
 to have things served up to him. 
 
 The next example, opening Washington National Airport to 
 jet traffic, is an interesting one because it illustrates how the 
 results of analysis are used and fed into a political process. You 
 may not know that Washington is served by three airports. 
 There's Washington National Airport, which is very close and 
 convenient to downtown Washington. Then there's Friendship 
 International Airport which is about 35 miles away in Maryland, 
 and Dulles International Airport which is about the same distance 
 away in Virginia. 
 
 With the coming of jet aircraft there was a great deal of 
 pressure on the Federal Aviation Agency which operated National 
 Airport to open it to jet traffic. The argument was from those 
 travelers who wanted the increased speed of jets with the 
 
 convenience of the downtown airport. Because some of the 
 travelers were also congressmen, it had great force. 
 
 Furthermore, the future equipment plans of the airline 
 operators who were serving Washington showed that in a couple 
 of cases, if National Airport wasn't open to jets they were just 
 going to quit serving it. They weren't going to have suitable 
 equipment. In some other cases, they might have continued 
 service with piston and turbo-prop aircraft but at a reduced level. 
 
 On the other hand, both Friendship and Dulles were built 
 to accept unrestricted jet traffic, both were large airports, 
 designed with the idea that there was going to be a big increase in 
 the volume of traffic, but the increase hadn't materialized at the 
 same time that traffic at National Airport was expanding to the 
 point of overcrowding. 
 
 There were some other factors, too. Obviously, the backers 
 of Friendship and Dulles airports were very strongly opposed to 
 any change in National Airport because it meant that the 
 development of their airports was going to be delayed. Dulles 
 International Airport, in particular, was a white elephant. It was 
 losing money every year. There were also people who would have 
 been in the jet patterns, that were very worried about the noise 
 increase and damage to their houses, as there was a lot of talk at 
 that time about sonic boom. All these groups were up in arms 
 about any change at National Airport. 
 
 Therefore, the Federal Aviation Agency decided that they'd 
 
 182 
 
THE CHOICE OF METHODOLOGY IN SOCIAL PROBLEM SOLVING 
 
 Figure 6. PREGNANCY TERMINATION YEARLY REPORT 
 MIC Simulation — Wayne County 
 
 Class and age 
 
 Community totals 
 
 Total Miscarriages Fetal deaths Births 
 
 Program patients 
 
 Total Miscarriages Fetal deaths Births 
 
 White 
 10-14. . , 
 15-19. . , 
 20-24. . . 
 25-29.. . 
 30-34. . . 
 35-39. . . 
 40-up... 
 
 Subtotal . . . 
 Percentage. 
 
 Nonwhite 
 
 10-14 
 
 15-19. „ 
 
 20-24 
 
 25-29 
 
 30-34 
 
 35-39 
 
 40-up 
 
 Subtotal . . . 
 Percentage. 
 
 Total 
 
 Percentage, 
 
 1 
 222 
 493 
 312 
 174 
 67 
 20 
 
 1,289 
 100 
 
 6 
 234 
 280 
 166 
 109 
 78 
 15 
 
 100 
 
 ',177 
 100 
 
 II 
 
 16 
 37 
 17 
 12 
 7 
 1 
 
 90 
 
 7.0 
 
 L 
 
 19 
 
 l'> 
 
 16 
 
 7 
 
 9 
 
 2 
 
 73 
 
 8.2 
 
 163 
 7.5 
 
 
 
 1 
 
 
 
 2 
 
 204 
 
 5 
 
 6 
 
 450 
 
 7 
 
 2 
 
 293 
 
 7 
 
 6 
 
 156 
 
 4 
 
 
 
 60 
 
 1 
 
 
 
 19 
 
 1 
 
 16 
 
 1,183 
 
 25 
 
 1.2 
 
 91.8 
 
 l 1.9 
 
 n 
 
 5 
 
 4 
 
 5 
 
 210 
 
 94 
 
 5 
 
 256 
 
 127 
 
 4 
 
 146 
 
 78 
 
 5 
 
 97 
 
 50 
 
 1 
 
 68 
 
 35 
 
 (i 
 
 13 
 
 7 
 
 20 
 
 795 
 
 395 
 
 2.3 
 
 89.5 
 
 l 44.5 
 
 36 
 
 1,978 
 
 420 
 
 1.7 
 
 90.9 
 
 ^.S 
 
 
 
 
 
 
 
 
 
 
 
 5 
 
 
 
 
 
 7 
 
 
 
 
 
 7 
 
 
 
 
 
 4 
 
 
 
 
 
 1 
 
 
 
 
 
 1 
 
 
 
 
 
 25 
 
 0.0 
 
 0.0 
 
 100.0 
 
 
 
 
 
 4 
 
 2 
 
 
 
 92 
 
 1 
 
 2 
 
 124 
 
 1 
 
 1 
 
 76 
 
 (i 
 
 4 
 
 46 
 
 ■' 
 
 1 
 
 32 
 
 
 
 
 
 7 
 
 6 
 
 8 
 
 381 
 
 1.5 
 
 2.0 
 
 96.5 
 
 6 
 
 8 
 
 406 
 
 1.4 
 
 1.9 
 
 96.7 
 
 This figure reflects the percent of the total population of pregnant women who participated in the clinic program as 
 realized in the simulation. 
 
 Figure 7. PROGRAM OPERATION YEARLY SUMMARY REPORT 
 MIC Simulation — Wayne County 
 
 
 
 Case load 
 
 
 Cost (dollars) 
 
 Clinic statistics 
 
 Total 
 
 Mean 
 per month 
 
 Standard 
 deviation 
 
 Total 
 
 Mean 
 per month 
 
 Standard 
 deviation 
 
 Maternal clinic: 
 
 388 
 
 2,840 
 
 63„3 
 
 359 
 1,115 
 
 317 
 1,356 
 
 131 
 
 215 
 
 365 
 
 6 
 
 48 
 
 
 
 14 
 
 29.8 
 218.5 
 
 27.6 
 85.8 
 
 24.4 
 104.3 
 
 10.1 
 
 16.5 
 
 28.1 
 0.5 
 3.7 
 
 4.1 
 24.1 
 
 5.3 
 
 22.4 
 
 5.3 
 11.0 
 
 3.1 
 
 3.9 
 
 5.3 
 0.5 
 2.4 
 
 I 70.125 
 
 \ 125,258 
 
 I 38,850 
 1,04 8 
 
 86 
 
 9,635 
 
 2,988 
 81 
 
 
 
 576.2 
 
 High risk cases, percent).. 
 Hospital deliveries : 
 
 
 
 2,47 2.0 
 
 Infant care clinic: 
 
 
 
 279.5 
 
 Physician office visits : 
 
 24 8 
 
 Family planning clinic: 
 
 
 Birth and death statistics 
 
 
 Still births „ 
 
 
 
 
 
 
 
 
 
 
 Effectiveness measures 
 
 Maternal clinic attendance ratio (scheduled visits/actual visits), 1.4 
 Total personnel utilization ratios (total hrs./hrs. allocated), Maternal, 
 Abnormality ratio (program abnormalities/population abnormalities), 1.6 
 
 0.7; 
 
 Infant. 0.5 
 
 183 
 
THE CHOICE OF METHODOLOGY IN SOCIAL PROBLEM SOLVING 
 
 have a study done to look at some of the factors, and that study 
 is what we're going to talk about. 
 
 It was intended to provide an assessment of the economic 
 and physical impact in the Washington, D.C., metropolitan area of 
 three policy alternatives. 
 
 One was excluding jet services at National Airport, and just 
 leaving things the way they were and seeing what happened. 
 
 The second was to permit short- and medium-range jet 
 
 aircraft operation at National Airport. Essentially, that meant 
 two- and three-engine jets, which were thought of in terms of a 
 500-mile limit for service. It turned out that Chicago, which a lot 
 of congressmen were interested in, was just beyond that, so it was 
 extended to include Chicago. 
 
 Finally, the third alternative was to permit all types of jet 
 aircraft to operate at National Airport. That meant allowing four- 
 engine jets in. 
 
 Figure 8. FACTORS AFFECTING WASHINGTON AREA AIR TRAVEL GROWTH AND 
 LOCAL AIRPORT SHARE OF PASSENGER AND AIR SERVICE ACTIVITY 
 
 EXTERNAL FACTORS 
 
 Growth of metropolitan areas 
 connected by air service to 
 Washington area 
 
 • Population, Employment, 
 Family Income 
 
 LOCAL FACTORS 
 
 • Growth of Washington-Baltimore Region 
 
 • Population, Employment, 
 
 • Family Income 
 
 • Changing Regional Development Patterns 
 Washington central business district vs. 
 surrounds 
 
 • Distribution of local points of origin 
 and destination of air travelers 
 
 Changing Local Transportation 
 
 Conditions 
 
 • Ground access time to three 
 Washington area airports 
 
 Forecast of 
 
 Washington Area Air 
 
 Travel Development 
 
 BEHAVIORAL FACTORS FOR AIR TRAVELING PUBLIC 
 Air travel potential as affected by: 
 
 • socioeconomic factors 
 
 • purpose of travel - business, personal 
 
 • travel cost 
 
 • travel time (including ground access time and speed 
 of air travel) 
 
 CHANGES IN AIR SERVICES 
 
 • type and quality of aircraft 
 
 • fare levels 
 
 • frequency of service 
 
 Figure 8 gives an idealized view of the real world system 
 that we're going to try to analyze. This is our abstraction of 
 what's going on. A couple of things that drive it are the shaded 
 boxes, the forecast of what's going to happen in travel develop- 
 ment. That's fed by a lot of external factors which you interpret 
 in the light of historical market development. Then there are 
 some behavioral factors for the air traveling public. How they're 
 going to make their choices of what airport they might go to, and 
 so on. Then there is the problem of the division of traffic among 
 the three airports, which depends on the changes in air services 
 and a number of other factors including airline decisions and 
 finally, all this ends up with FAA decisions. 
 
 The analysis model that was used is shown schematically in 
 figure 9. In order to provide some kind of time horizon for each 
 of the two alternatives involving jet service, the study undertook 
 to estimate and project for the years 1966 through 1980: (1) the 
 net economic benefits or costs to users, to operators of aircraft 
 services, and the community, (2) the net financial gain or loss to 
 FAA or the operator of National Airport, Dulles International 
 Airport, Friendship International Airport and the airway system, 
 and (3) the new physical demands for new and additional airport 
 and airway facilities. 
 
 If you let jets in and traffic expands you'd have to build 
 more and bigger runways, expand the terminal, change the fueling 
 
 184 
 
THE CHOICE OF METHODOLOGY IN SOCIAL PROBLEM SOLVING 
 
 facilities,, provide more parking facilities, and meet all of the 
 investment demands. 
 
 Looking at figure 9 the logic started out in the upper 
 left-hand box. For each of 100 metropolitan areas which were 
 connected to Washington by air service, the socioeconomic 
 composition of the area was projected over the time horizon of 
 the study. Then, using trip generation factors, business travel by 
 industry group was projected and, depending on socioeconomic 
 factors, the number of trips for personal travel was generated. 
 These factors came from surveys of air traffic. It was indicated 
 that wholesale and retail trade would generate the equivalent of 
 1.78 business trips for each of its employees, construction would 
 generate .92 trips per employee, amusement and recreation would 
 generate .48 business trips per employee per year, and so on. 
 Using all these factors and projecting the composition of the area, 
 it was possible to generate how much traffic might be expected 
 over this time horizon. 
 
 There was a separation between flights that terminated in 
 Washington (or had interchange in Washington) and through- 
 flights. Through-flights were subject to different kinds of con- 
 siderations than those terminating, particularly for the model 
 deciding whether people would go to one airport or the other. 
 
 This split model, or model for splitting the passenger traffic 
 among the airports is a function of the total travel time. Total 
 time includes the time it takes you to get to the airport, the delay 
 time in how long you'd have to wait for the next available flight 
 depending on which airport you go to, and travel on the other 
 side. 
 
 To compute total time, it was necessary to compute ground 
 travel time to the airport. To do this, population growth in the 
 Washington area was projected by zones. With the computation 
 of trips and traffic for each city paired with Washington, block of 
 time during the day, and original destination location within the 
 Washington area, it was possible to project the split of the traffic 
 and estimate how much would go to each of the three airports. 
 
 Finally, all this traffic generated requirements for airport 
 capacity as shown on the right-hand side of the output measures. 
 There was an imputed value or benefit to users in terms of how 
 their time was to be valued. There were some conventions among 
 transportation analysts as to how you value a traveler's time. In 
 automobile traffic it's valued at about 4 cents a minute, air travel 
 is at a much higher rate, because it has a lot of business travel by 
 people who have higher paying jobs, and their time is more 
 valuable. 
 
 Figure 9. WASHINGTON AIRPORT ANALYSIS MODEL 
 
 INPUTS 
 
 DEMAND MODEL PARAMETERS 
 
 • Average family income 
 
 • Employment 
 
 • Population 
 
 • Base period traffic for each 
 O ■ D internal zone, external 
 city 
 
 AIRCRAFT PARAMETERS 
 
 • Total passengers/yr. per 
 aircraft type 
 
 • Range constraints per 
 aircraft type 
 
 a Alternative rules for WNA 
 
 TRAVEL TIME PARAMETERS 
 
 • Ground travel time zone- 
 airport 
 
 • Airport-city block time 
 each aircraft type 
 
 • Ground travel time 
 external city 
 
 Compute total traffic for 
 new period between 
 zone-city pairs 
 
 Compute aircraft flight 
 schedules for each 
 airport-city pair and 
 traffic split by zone- 
 airport-city combinations 2. 
 
 Compute average total 
 travel time for each 
 zone-airport-city 
 combination 
 
 Computed adjusted total 
 passenger volume for 
 new schedule frequencies, 
 each zone-airport-city 
 combination j_ 
 
 Total adjusted passenger 
 volume for zone-city pair 
 for previous period, split 
 by business-personal- 
 directional components g, 
 
 Reiterate for next 
 period computation 
 
 Total adjusted passenger 
 volume for zone-city 
 pairs, previous period 
 
 Number aircraft schedules 
 by type, airport-city 
 combination 
 
 Total-all aircraft 
 operations each airport 
 
 Compute total travel time 
 savings for WNA alternatives 
 2&3 over alternative 1 
 
 Add all other passenger 
 traffic components 
 each airport 
 
 ■J 
 
 OUTPUTS 
 
 Noise 
 
 Evaluation 
 
 (Aircraft 
 Activity 
 Mpacnrp 
 
 I Passenger 
 Activity 
 Measu re 
 
 Noise Contour 
 
 Maps 
 
 (BB&N Method) 
 10. 
 
 Airport 
 Capacity 
 
 Analysis 
 
 x Imputed 
 
 Total 
 
 
 value per 
 
 Public 
 
 
 traveler 
 
 Benefit $ 
 
 
 hour 
 
 
 13. 
 
 Estimate cost 
 revenues and 
 not revenue, 
 each airport 
 
 185 
 
THE CHOICE OF METHODOLOGY IN SOCIAL PROBLEM SOLVING 
 
 Further, it was possible to estimate the costs and the net 
 revenues to each airport because as the traffic changed so the 
 pattern of concessions and fees would change. 
 
 In addition to what was done with the scheduled airlines, it 
 was necessary to generate what was going to happen to general 
 aviation. That is, how much flying would be done by people who 
 fly their own planes. 
 
 Lastly, shown in box number 10, the noise contours were 
 patterned. That analysis was done by a special group who are the 
 leading experts on noise analysis in the country, possibly the 
 world. The FAA had run some tests in Oklahoma City because 
 there was a lot of concern about sonic boom and the damage 
 from sonic boom. There was no doubt that there was going to be 
 some increase in noise, and it was a question of how much, how 
 objectionable it would be, and what was going to happen. 
 
 Roughly, the upshot of all this was that the study found 
 that there would be some diversion of flights from Friendship and 
 Dulles, and the effect of that would be to delay the development 
 of traffic at those airports. But there were strong overall 
 economic benefits to opening National Airport to short- and 
 medium-range jets. 
 
 After some pulling and hauling among the people 
 involved— including a court suit— FAA adopted a policy of 
 permitting short- and medium-range jets, two- and three-engine 
 jets, in National Airport. 
 
 A little later, because there was so much complaint about 
 the noise, there was an embargo on any jet flights after 10 p.m. 
 So if you're on a jet flight that's going to land in Washington after 
 10 in the evening, you're not going to land at National Airport. 
 
 Mr. Barabba: And you can rest assured that you're going to come 
 in over the Potomac River, no matter what the weather is. Was 
 this put into a computer model? 
 
 Mr. Nisselson: Yes. It was a big computer exercise. 
 
 One thing that's clear about this problem is that it's going 
 to come up again. But now there are going to be different 
 answers. There will be a rapid transit line out to Dulles which will 
 change them. Then there's going to be a big airport built out 
 further with heliport service. 
 
 This study cost about $96,000, and it took 6 months. One 
 of the things that happened is that people began raising questions, 
 and questions were suggested by analysis that hadn't been raised 
 before. After the bulk of the work was done there continued to 
 be a little further analysis. 
 
 Student: Would you make a bid before you do the study? 
 
 Mr. Nisselson: Yes. A typical process by which studies like this 
 get done is that an agency has an idea of the study they want. 
 They should know something about the state of the art, and if 
 they're in transportation they know these kinds of studies have 
 been done, and maybe they're expecting some new wrinkles to be 
 put in. But they'll write up what they want done, the problem 
 that they want solved, and what kinds of answers they want, and 
 they'll put out what's called a request for proposal. 
 
 They might receive 50 or 100 proposals from companies 
 that say they're going to do it, how smart they're going to be, and 
 all the great things they're going to do. Typically, there's a 
 technical evaluation on the part of the government agency to 
 identify the top competitors who will be called in for further 
 interview to see how credible they are. Among the satisfactory 
 
 186 
 
 competitors, generally a decision is made on the basis of low 
 price. If a government agency, at least in the Federal government, 
 doesn't pick the lowest bidder it must justify why. Procurement 
 regulations are a little less stringent at State and local levels. In 
 fact, many local communities don't have to use competitive 
 bidding at all. When you get bids you may find out you 
 underestimated the problem. But generally, an experienced 
 agency will indicate about how much they're prepared to spend. 
 You don't start out getting studies like this when you don't know 
 anything about the problem or what some of the ways of dealing 
 with it are. 
 
 I couldn't resist this last example which is an example of a 
 highway design decision, and it illustrates a couple of things that 
 interest me personally. One is that the timing of a decision may 
 become a policy question. Figure 1 shows, on the "x" axis along 
 the bottom of the chart, some hypothetical actual average daily 
 vehicle traffic per hour, 1,500, 2,800, 4,800, and 6,800. Suppose 
 you make a four-lane forecast and the traffic is, say, only 1,500 
 vehicles per hour, then you have overbuilt the highway. Plus, you 
 have an investment cost, because you have invested your money 
 in building this highway when you could have done something 
 else with the money. That cost continues until the traffic reaches 
 a four-lane forecast when there is no further investment cost. If 
 you underestimate the traffic and it keeps going up, then that 
 first dotted curve going up exponentially shows what the cost of 
 underdesign is; that cost arises from congestion and its conse- 
 quences. Similarly, a six- and eight-lane forecast is shown. What 
 this chart illustrates is that for a highway construction program, 
 forecasting errors incur costs through incorrect design decisions. 
 
 It also shows, incidentally, that the costs of overdesign and 
 underdesign are not necessarily symmetrical. There's not the same 
 penalty for being off by a given amount of traffic. This figure 
 suggests that to the extent that an improved data base could lead 
 to a reduction in forecasting errors, there is a payoff for good 
 data. In fact, an improvement in the data base is represented by 
 the reduction of cost that would be due to the reduction of 
 forecasting errors. If you really understand a system as well as 
 people understand highway design, you can put a figure on how 
 much information is worth, which is a situation we're not very 
 often in. 
 
 Turning to the policy issue, it is interesting to observe that 
 the costs of underdesign are largely private costs incurred by the 
 individual driver and passengers. These are the costs of the 
 increased time spent in travel because of congestion, and there is 
 an increase in the vehicle operation cost due to congestion. 
 
 On the other hand, the cost of overdesign is a public cost 
 which is relatively highly visible through tax loads. My feeling is 
 that this may be part of the explanation of why highway facilities 
 at the local level tend to be underbuilt. I don't know if it's been 
 your experience, but in the Washington area, at least, they are 
 always building facilities behind the demand, and when a new 
 bridge is put across the Potomac River or a new highway is 
 completed, as soon as it's completed, bang— it's overused and 
 congested. Traffic engineers can't be that short-sighted. I think 
 it's really this policy issue that's involved,— where the costs are 
 and who bears them. Remember, a lot of decisions have a time 
 horizon, and the question of how good a decision is may be 
 related to a point in time. It's good at one point in time; it's not 
 so good at another point in time. Again, as a statistician and 
 looking at the question of data and the value of data, it's an 
 interesting example where you can really put a price on 
 information. 
 
THE CHOICE OF METHODOLOGY IN SOCIAL PROBLEM SOLVING 
 
 Figure 10. CONDITIONAL ERROR COST FUNCTIONS OR OPPORTUNITY 
 LOSS FUNCTIONS ACCRUING FROM FORECASTING ERRORS 
 
 1000 
 
 900 
 
 800 
 
 700 
 
 600 
 
 £ 500 
 
 400 
 
 = ■3 300 
 
 200 
 
 "D 
 
 o 
 
 CD 
 N 
 
 "D 
 
 
 M— 
 
 CD 
 
 o 
 
 CD 
 
 I/) 
 
 
 "D 
 
 T5 
 
 C 
 
 r 
 
 CD 
 
 CD 
 
 3 
 
 > 
 
 O 
 
 +-* 
 
 ,.r 
 
 o 
 
 ■t-j 
 
 CD 
 
 r 
 
 a 
 
 
 CD 
 
 +-' 
 
 O 
 
 i/i 
 
 
 O 
 
 c 
 
 o 
 
 CD 
 
 
 
 1_ 
 
 V> 
 
 n 
 
 a> 
 
 ;=. 
 
 "O 
 
 i_ 
 
 
 OJ 
 
 c 
 
 OS 
 
 
 r- 
 
 O) 
 
 
 U 
 
 +- 1 
 
 c 
 
 CD 
 
 (-0 
 
 CD 
 U 
 
 CD 
 
 <D 
 
 E 
 
 o 
 
 M— 
 
 ._, 
 
 o 
 
 M— 
 
 ■I-' 
 
 
 
 
 
 (D 
 
 a) 
 
 n 
 
 r^ 
 
 T2 
 
 b 
 
 O 
 
 ■^ 
 
 "- 
 
 (^ 
 
 M— 
 
 o 
 
 CD 
 
 O 
 
 *-• 
 
 100 - 
 
 Cost per mile of 
 unutilized capacity 
 (overestimated demand) 
 
 Cost per mile to users 
 due to congestion 
 (underestimated demand) 
 
 I 4 lane - Forecast 
 | of average daily 
 2800vehicle/hr. 
 
 8 lane forecast • 
 6 lane forecast 
 4 lane forecast « 
 
 6 lane - Forecast 
 of average daily 
 4800 vehicle/hr. 
 
 I 
 
 I 
 
 8 lane - Forecast 
 of average daily 
 6800 vehicle/hr. 
 
 1500 2800 4800 
 
 Actual Average, Daily Vehicle Traffic Per Hour 
 
 6800 
 
 Student: How can I tie opportunity loss to interpreting data? 
 
 Mr. Nisselson: The term "opportunity" just means what else you 
 could do with the money. It's related mainly to the fact that if 
 you've overbuilt you're sinking money into the ground, and you 
 can't do anything with it. 
 
 Student: It looks to me, on this chart, that it would be safer to 
 overbuild. 
 
 Mr. Nisselson: It depends on whether you're the guy who says 
 taxes are going to be raised, and you're not going to be elected. 
 
 Student: It looks like it's enormously safer to overbuild. Am I 
 misinterpreting it? 
 
 Mr. Nisselson: No, you're not misinterpreting it. But on the other 
 hand, people are sitting around in their cars, and their time is 
 worth 4M> cents a minute. Who pays attention to costs like that? 
 
 187 
 
THE CHOICE OF METHODOLOGY IN SOCIAL PROBLEM SOLVING 
 
 Student: On this cost per mile of inbalance on the "y" axis, is 
 that per year? 
 
 Mr. Nisselson: It's per year. Those costs were based on some work 
 that Professor John Meyers did when he was at the Harvard 
 Transportation Research group before he became the head of the 
 National Bureau of Economic Research. 
 
 Mr. Barabba: I'd like to say a little more about data and the value 
 of good data. Models for analyzing complex systems consume 
 enormous amounts of data, and the results are not likely to be 
 much better than the quality of the data that went into them. It 
 is generally a mistake to assume that the power of analytic tools 
 can compensate sufficiently for inadequacies in the input data. 
 And that is one tie between our subject this morning and the 
 Bureau of the Census. The Bureau's censuses and surveys are an 
 enormous resource for data, and particularly small area data, 
 obtained on a uniform basis over the entire country. And the 
 Census Bureau goes to great lengths to try to assure that the data 
 produced are of a known quality and to inform users of its data 
 availability and the limitations or qualifications. 
 
 Besides its periodic censuses and surveys, the Bureau also 
 provides other current data: for example, the cooperative 
 Federal-State program which produces annual estimates of the 
 population by State and county; the program for the Office of 
 Revenue Sharing which produces estimates of population and per 
 capita income for 39,000 places eligible under revenue-sharing; 
 the cooperative program with the Social Security Administration 
 which produces the publication County Business Patterns that 
 gives number of establishments by county with detail by kind of 
 business and size measured by number of employees. Data such as 
 these are useful in themselves, and for developing relationships 
 for model building, but they are also useful as benchmarks. For 
 example, a research group at the Chamber of Commerce in 
 Chicago has been testing out a method for producing current 
 estimates of population by census tract, using data on telephone 
 connections specially compiled for them by Illinois Bell. The data 
 are adjusted to conform to the estimates for Chicago and the 
 balance of Cook County taken from the Federal-State coopera- 
 tive population estimates program I mentioned. They are very 
 pleased with their first results. 
 
 In addition to data, the Bureau also is active in developing 
 methods for relating census data to local data and, in particular, 
 to relating different data sets on the basis of geographic location 
 within a local area. This is a special concern of the Census Use 
 Study. Major activities in developing small area estimates are of 
 two kinds. One approach is called Synthetic Estimates. This starts 
 with a larger area for which reliable relevant statistics are 
 
 available, say from a sample survey, but for which the corre- 
 sponding estimates from the sample can't be made for small areas 
 within the larger area with a high enough reliability because there 
 isn't enough sample in the small areas. To develop a so-called 
 synthetic estimate the units, say, population or housing units, are 
 tabulated for the larger area and the smaller area into cells 
 defined by a number of variables related to the statistic of 
 interest. Suppose you are interested in estimating the number of 
 units by small area having a given characteristic. You would take 
 the percent of units having the characteristic in the larger area in 
 each cell and apply that percent to the number of units in that 
 cell for each of the small areas. In the 1970 census this method 
 was used to develop estimates of dilapidated units with all 
 plumbing facilities for States, counties, SMSA's and places of 
 1 0,000 population or more. The method assumes that the percent 
 of units in a given tabulation cell having the characteristic of 
 interest is the same in all subareas. When this assumption is not 
 true the synthetic estimate is biased. It is not possible to estimate 
 the bias for any particular area separately, but in the method as 
 used by the Census Bureau it was possible to develop an estimate 
 of a root mean square error for groupings of publication areas, or 
 areas for which data were published separately. 
 
 The other kind of activity is based on more conventional 
 regression estimates such as those used in making current 
 population estimates. Such regressions attempt to relate popula- 
 tion change, which, of course, is unknown unless a special census 
 has been taken, to so-called indicator variables or changes in such 
 variables. By indicator variables we mean characteristics of the 
 area which are correlated with the variable of interest (population 
 change). For example, in the case of estimates by census tract in 
 Chicago I mentioned, the indicator variable was telephone 
 connections. The methods now being developed attempt to 
 incorporate into the regression equations independent estimates 
 of the variable of interest for a sample of areas which may be 
 based on either a census or sample survey of the sample areas. 
 This approach, which seems promising, is applicable to any area 
 comprised of subareas and in which independent estimates are 
 available for a sample of areas.The term "sample" needs comment. 
 Just any subset is not a sample, but the subset should be selected 
 as a probability sample of all areas for which the regression 
 equation is to be developed. It is a matter for real concern when 
 indicator variables come from administrative records, or program 
 records, whether the data are good enough. It's absolutely 
 necessary if you want to get a good data base from administrative 
 records and maintain a consistent base from year to year to have 
 a quality control operation at the source. This is particularly 
 important for social indicators, and for local uses, in which data 
 may come from programs in which the focus of attention is the 
 program rather than the program statistics. 
 
 188 
 
GBF/DIME SYSTEM-THE DIFFUSION OF AN INNOVATION 
 
 Vincent P. Barabba, Director, U.S. Bureau of the Census 
 
 "I would like to emphasize that in the adoption of many systems it is crucial to 
 ensure the ultimate choice not be limited to only the nominal head or a relatively few 
 leaders of an organization. Those employees in any position of influence, as well as those 
 who would use the system, should be alerted to the impending decision and encouraged 
 to express their opinions." 
 
 Vincent P. Barabba 
 
 Mr. Barabba: The past sessions have described systems, tools, and 
 information available as an outgrowth of Federal or private 
 statistical efforts. However, we still need to explore the process 
 and problems of having one of these sytems or tools adopted by 
 organizations or individuals. 
 
 For the purpose of this discussion I would like to use the 
 Census Bureau's GBF/DIME system and the history of its 
 adoption as the case in point. The GBF/DIME system (geographic 
 base file/dual independent map encoding) provides a more 
 systematic utilization of census data and administrative records 
 for program evaluation than was heretofore available to statis- 
 ticians, planners and organizations interested in the analysis of 
 small area data (see appendix). At one time the GBF/DIME sys- 
 tem was an entirely new statistical tool being offered to a large 
 public of potential adopters, many of them totally unaware of its 
 potential uses. A lack of understanding of the system was 
 partially responsible for an initial reluctance on the part of some 
 to accept the system. Today the GBF/DIME system is a generally 
 accepted tool employed throughout the U.S. This will not be a 
 discussion of why a community or organization should have 
 adopted or should now adopt the GBF/DIME system. Instead, I 
 want to orient my presentation so that we can outline the prob- 
 lems and procedural steps involved in having any new system or 
 tool adopted. 
 
 Everett Rogers, Professor at the University of Michigan, 
 clearly illustrates that an extremely meaningful concept which is 
 known by, and beneficial to, individuals or those persons 
 representing an organization may not be accepted because of a 
 failure on the part of those presenting the new tool to properly 
 understand and utilize the innovation adoption process. 
 
 For the purpose of our discussion let me identify an 
 innovation as being any idea, way of doing things, or material 
 object that is new to a given individual or group of individuals. 
 Therefore, an innovation could be widely used by some people, 
 and yet new to a given group. The innovator may be said to be an 
 individual who promotes the adoption of the new idea or tool. 
 Almost every area of public concern has one or more types of 
 innovators or concept promoters. Some examples are technical 
 assistance workers in less developed countries, dealers in new 
 products, salesmen, public health officials, doctors, and adminis- 
 trators, among many others. 
 
 Now, conversion to a computerized geographic file system 
 is not much different than converting to any computer-oriented 
 system. Recently, an officer of a large corporation identified the 
 
 1 Everett Rogers, Communication Of Innovations, The Free Press, New 
 York, New York, 1971; Diffusion Of Innovations, The Free Press, New 
 York, New York, 1962. 
 
 principal reasons why firms are slow to adopt new computer 
 techniques in their marketing programs. A very real difference 
 between the marketing and computer systems personnel, and the 
 organizational separation of these units, contributes to misunder- 
 standing and lack of communication; inflated expectations of 
 computer capabilities are created by overselling, which often lead 
 to misuse of the computer by top level management. He added 
 that competition discourages innovations because of the time 
 needed to make the system operational and relatively error-free. 
 
 Likewise, persuading a community or organization to adopt 
 a computerized information system is not easy. Critics warn that 
 in implementing the system, the local planners will soon find they 
 cannot afford it; their first program will be unsuccessful; they will 
 be personally embarrassed; they will try to save the sinking ship, 
 and 50 percent of what they finally develop will be untimely, 
 unnecessary, and too expensive. 
 
 Therefore, before you introduce any innovation to an 
 agency or corporation, you must really understand the people 
 involved, both as individuals and as a group. To understand them 
 as individuals, we need to know several things. What is their 
 ability to handle sophisticated tools? Do they basically under- 
 stand what's involved? How secure are they in their positions? 
 What are their values? And, can they influence others? 
 
 To understand how the members interact, we need to know 
 how traditional the group is. Do they do something a certain way 
 because they have always done it like that, or are they willing to 
 try new ideas? What are the organization's characteristics? The 
 pressures developed by each organization must be handled 
 differently. It is equally important to remember that the 
 acceptance of an innovation does not ensure its use by the 
 organization. Therefore, the first step is to take a good look at 
 the organizational environment from both points of view. 
 
 Rogers cites four phases in the innovation adoption process. 
 These are the knowledge, persuasion, decision, and confirmation 
 stages. 
 
 First, in the knowledge stage the individual is exposed to 
 the innovation's existence, and he gains some understanding of its 
 functions. The individual's attitude toward the innovation hinges 
 on whether he sees the innovation as a solution to a problem or a 
 threat to his own continued employment. 
 
 Second, during the persuasion stage, the individual forms a 
 favorable or unfavorable opinion about the concept. The indi- 
 vidual becomes more psychologically involved as his knowledge 
 of the innovation increases. Since there is some degree of risk 
 involved in adopting an innovation, he may seek reinforcement of 
 his attitude from his peers. 
 
 During the third phase, the decision stage, the individual 
 undertakes activities which will ultimately lead to the adoption or 
 
 189 
 
GBF/DIME SYSTEM-THE DIFFUSION OF AN INNOVATION 
 
 PARADIGM OF THE INNOVATION-DECISION PROCESS 
 
 (ANTECEDENTS) 
 
 (PROCESS) 
 
 (CONSEQUENCES)* 
 
 Receiver Variables 
 
 Communication Sources 
 
 1. Personality 
 characteristics 
 (e.g., general 
 attitude toward 
 change) 
 
 2. Social 
 characteristics 
 
 (e.g., cosmopolitism) 
 
 3. Perceived need for 
 the innovation 
 
 4. Et cetera 
 
 (Channels) 
 
 KNOWLEDGE 
 I 
 
 PERSUASION 
 II 
 
 DECISION 
 III 
 
 Social System Variables 
 
 Perceived Characteristics of Innovations 
 
 1. Social System Norms 
 
 2. Tolerance of Deviancy 
 
 3. Communication Integration 
 
 4. Etcetera 
 
 1. Relative Advantage 
 
 2. Compatibility 
 
 3. Complexity 
 
 4. Trialability 
 
 5. Observability 
 
 Adoption 
 
 Continued Adoption 
 
 Discontinuance 
 
 1. Replacement 
 
 2. Disenchantment 
 
 CONFIRMATION 
 IV 
 
 Rejection 
 
 Later Adoption 
 
 Continued Rejection 
 
 ■TIME- 
 
 * For the sake of simplicity we have not shown the consequences of the innovation in this paradigm but only the consequences of the process 
 
 rejection of the innovation. Rogers points out that most 
 individuals do no accept an innovation without first using it on a 
 probationary basis to determine its utility. Because of the obvious 
 advantages of a trial prior to adoption, those innovations available 
 for a trial period are generally adopted more rapidly. But, it is 
 just as possible the trial will result in the rejection of the concept. 
 Rogers suggests that while rejection may occur at any time, it 
 usually happens at this point in the adoption process because the 
 results of trail are misinterpreted. 
 
 An example of this misinterpretation comes from the 
 report of an Iowa farmer who tried to increase his crop yield by 
 applying nitrogen fertilizer. Although he gained a higher crop 
 yield, the farmer erroneously attributed the death of two cows, 
 that had actually died from eating wet clover, to the new 
 fertilizer. Therefore, he rejected its use. Later interviews indicated 
 the farmer's neighbors were not using the nitrogen fertilizer 
 because of his experience. 
 
 Based on empirical evidence, Rogers and others suggest 
 there is an additional stage beyond the initial decision to adopt or 
 reject the innovation. This fourth stage is one of further testing 
 by direct application and confirmation. The individual seeks 
 further reinforcement of his attitude. 
 
 I would like to emphasize that in the adoption of many 
 systems it is critical to ensure the ultimate choice not be limited 
 to only the nominal head or a relatively few leaders of an 
 organization. Those employees in any position of influence, as 
 well as those who would use the system, should be alerted to the 
 impending decision and encouraged to express their opinions. 
 
 An important ingredient dictating how an innovation fares 
 through this process of adoption is the innovation itself. But, how 
 the concerned individuals perceive an innovation also affects, the 
 adoption. The individual's perception of the innovation centers 
 on five characteristics: (1) the ability to identify the relative 
 advantages of having such a system versus not having it; (2) the 
 ability to illustrate the manner in which such an information 
 system is compatible to present organizational procedures; (3) the 
 ability of successful and qualified people from many areas of 
 endeavor to quickly absorb the technical jargon of a vast and 
 complex tool; (4) the need to divide the concept and test it in a 
 way that no person or organization feels threatened; and (5) 
 communicating the results of the effects of such an information 
 system. 
 
 Depending on the individual adopters the use of the Census 
 Bureau's community geographic-oriented information system has 
 not had universal acceptance because of the reactions to the 
 characteristics I have just mentioned. 
 
 In assessing the adoption of the Bureau's GBF/DIME 
 system, three groups emerge: Those who had almost instanta- 
 neous support for the concept, those who were neutral, and those 
 who were opposed. 
 
 Supporters of the system can be subdivided by their 
 motives for support. Realizing the need for new techniques to 
 structure and retrieve the vast amount of data generated by local 
 agencies, and the difficulty of integrating the available data for 
 small geographic areas, some supporters recognized the potential 
 of the concept to make their work easier and more accurate. 
 
 190 
 
GBF/DIME SYSTEM-THE DIFFUSION OF AN INNOVATION 
 
 These were primarily planners and researchers. Another subgroup 
 was moved to support the concept because of the Bureau's 
 sponsorship. The Bureau's reputation in statistical and geographic 
 matters is widely known and respected. These supporters were 
 accustomed to working with the Bureau, and they believed that 
 adoption of the concept would result in better censuses. They 
 also realized the capabilities of the system have been successfully 
 demonstrated in New Haven, Conn., and Los Angeles, Calif., and 
 the documented applications and procedural guidelines are 
 readily available to them. 
 
 Those individuals who were neutral, such as planning 
 directors and county department heads, were not likely to be the 
 direct users of the concept. This group would be using the 
 product of the system, rather than the system itself, and they 
 tended to rely on the advice of technical subordinates. 
 
 Those who were opposed can also be subdivided by their 
 motive. One subgroup, agencies that already had geographic base 
 files and address matching software, perceived the concept as a 
 threat to their operations. They felt its adoption would detract 
 from their systems or lead to a centralization that would 
 eventually result in a transfer of functions over which they had 
 control. Others not immediately supporting the idea were budget 
 analysts and administrators, whose job it was to question all new 
 programs and projects. 
 
 Now that we have examined the basic elements of the 
 adoption process and outlined those groups that were exposed to 
 the GBF/DIME system, let's view the adoption of the system 
 through Roger's four stage process. 
 
 Knowledge 
 
 The identification or matching of data to geographic units 
 is not new. In the past, before electronic data processing, the 
 geographic identification was done manually and crudely, because 
 the sources of geographic identification were mainly maps or city 
 directories. For example, the numerous transportation studies 
 which started in the late 1940's identified the origin and 
 destination of trips in order to code them to a transportation zone. 
 In the 1955 transportation study of Washington, D.C., approx- 
 imately 10 to 15 percent of the households were sampled, and 
 about 180,000 origins and destinations were located and iden- 
 tified. This was done using Sanborn maps and Polk and other 
 directories. 
 
 The first exposure of governmental agencies and private 
 organizations to a standardized type of geographic file adaptable 
 to a computer information system was the Address Coding Guide 
 (ACG) which was the forerunner of the GBF/DIME system 
 prepared for use in the 1970 Decennial Census of Population and 
 Housing. However, these files were developed to meet a specific 
 Bureau need, and the idea that they could have potential use or 
 value at the local level was not widely comprehended outside the 
 Census Bureau. 
 
 As the potential for, and the GBF/DIME system itself 
 evolved, information regarding the system's capabilities was 
 distributed through meetings and conferences sponsored by the 
 Bureau, published papers, the Bureau's SAD (small area data) 
 Notes, and professional society meetings. The Bureau's Geo- 
 graphy Division, because of its continuing contacts with local, 
 State, and Federal agencies, handled the development and 
 implementation of the file. The Census Use Study was assigned 
 the task of identifying research applications of the file to local 
 needs. 
 
 Persuasion 
 
 The persuasion phase began in the early 1970's, and the 
 Bureau immediately faced some hard questions from the 
 potential users. By 1972, governmental agencies and private 
 organizations had some knowledge of the system, but there was 
 confusion about the differences between the ACG, GBF, and 
 DIME programs and their capabilities. Potential users were 
 uncertain which system best filled their needs. Other interested 
 parties were concerned with the availability of support programs; 
 they were uncertain of the costs and origin of financing to 
 develop user programs for the file. In other words, there were the 
 costs of developing the file, and the costs of establishing programs 
 to use with the file. A third consideration was the maintenance 
 and extension of the files. The Bureau's CUE (Correction, 
 Update, and Extension) program was oriented to several updates 
 at 2- or 3-year intervals— funds permitting— to achieve the most 
 accurate and current file of all areas for use in coding the 1980 
 census address file, 
 
 There were many programs for which a file and a 
 computerized geocoding system could be used, such as those of 
 Housing and Urban Development, (HUD), the Office of Educa- 
 tion, and the Law Enforcement Assistance Administration 
 (LEAA). Each agency that worked with the Census Bureau in the 
 GBF/DIME system development received a detailed report of the 
 concept capabilities. However, agency administrators were neutral 
 toward the new system. They foresaw the benefits of the system, 
 but would become involved only if additional monies were made 
 available to finance its implementation. Fortunately, organiza- 
 tions such as HUD and the LEAA made funding available. 
 
 The GBF/DIME system was geared to standardization on a 
 nationwide basis, meaning that techniques developed for one area 
 would become available to other areas in a form which could be 
 readily utilized. The CUE procedures were designed by the 
 Bureau, thus eliminating development costs for the local agencies, 
 which were offered free service to process the sytem, if they had 
 no computers of their own. 
 
 Of course, the benefits of standardization required all local 
 agencies to conform to common formats and data elements, at 
 least in that portion of the file used by the Bureau. This 
 conformity required some agencies to earry data which were of 
 no apparent use locally. 
 
 Despite the advantages of the system, the preliminary 
 criticisms of those initial users reinforced the negative attitudes of 
 potential users. The criticisms concerned the "overselling" of the 
 system's availability and its immediate application and use, and 
 the imperfections in the files. Admittedly, those first files were 
 not perfect. Beyond the anticipated problems, there were 
 difficulties arising from the poor quality of local address 
 information, and the lack of a corrected, updated system and an 
 edit and correction process. There was a lack of support from 
 local officials, who failed to recognize the value of this sytem as a 
 management information and decisionmaking tool capable of 
 relating local data to the geographic distribution of socio- 
 economic data. Because of the criticism, problems, and lack of 
 support, some administrators adopted alternate systems to meet 
 their needs rather than wait for the perfected GBF/DIME system. 
 
 To combat the bad publicity and administrative reluctance 
 to adopt the system, the Bureau aimed its persuasion efforts at 
 the agencies' mid-level technicians in the hope of influencing the 
 decisionmakers through their staffs, 'n these mid-level meetings, 
 the merits of competing systems were debated. For example, one 
 existing system was more efficient than the GBF/DIME system. 
 But the latter provided more flexibility than the existing system. 
 
 191 
 
GBF/DIME SYSTEM-THE DIFFUSION OF AN INNOVATION 
 
 Decision 
 
 As I stated earlier, the decision stage is when an individual 
 actually engages in activities, such as a trial use, which lead to the 
 adoption or rejection of the innovation. In our case, meeting with 
 Bureau representatives may have been sufficient for some per- 
 sons; for others, a written agreement was necessary. The Bureau 
 held a number of conferences, meetings, and workshops at the 
 local and national level for those individuals interested in develop- 
 ing the system or already using it. 
 
 The GBF/DIME system has a wide variety of uses, as evi- 
 denced by the comments and testimony of those already employ- 
 ing the system. Operating units throughout the country have 
 adopted the geocoding and mapping technique to provide inven- 
 tories of everything from streets and roadways to water and sewer 
 systems, from motor vehicle registration to voter registration, and 
 for planning the best location of firehouses, day care centers, 
 schools, and public recreation sites. Schedules of social service 
 caseworkers have been plotted for greater efficiency. Mapping 
 and evaluating public transit services and profiling neighborhoods 
 for their crime, accident, and redevelopment characteristics have 
 all been undertaken. With the system's assistance residency 
 
 checks and tax collection evaluations are performed. And at least 
 one newspaper has employed the system to evaluate its subscrip- 
 tion campaign and assist its advertisers. 
 
 At this point the many uses should be obvious. Naturally, 
 there are undiscovered applications as well. As the system's 
 adopters become aware of the already developed uses, its utiliza- 
 tion will increase, or to put it another way, the decision to adopt 
 the system will be affirmed and reaffirmed. 
 
 Confirmation 
 
 There are now 205 SMSA's that have prepared a file and 
 have begun using the system, as their various local management 
 operations and research programs permit. Not all 205 files are in 
 the same condition. Even after the initial development of a file, 
 there are naturally residual errors that must be corrected. In 
 addition, characteristics of urban growth or change must be in- 
 cluded in the files. 
 
 Sixty-five SMSA's have yet to prepare a file, although a 
 large number of these will do so as soon as money and personnel 
 become available. 
 
 Appendix. Geographic Base (DIME) System 
 
 Background 
 
 In 1970 the Bureau of the Census conducted the Nine- 
 teenth Decennial Census of Population and Housing by a 
 combination of two methods: a mail canvass in the large urban 
 areas of the nation and a house-to-house enumeration in the 
 remainder of the country. 
 
 For the urban cores, covering the postal city delivery areas 
 of 145 standard metropolitan statistical areas (SMSA's) in which 
 the mail-out/mail-back procedures were used, a method was 
 needed to code individual addresses on the mailing list to specific 
 geographic areas for tabulation purposes. A master coding file was 
 developed to code approximately 40 million addresses by 
 computer to the appropriate geographic areas. The coding file 
 developed specifically for this purpose was called the address 
 coding guide (ACG). 
 
 The address coding guide worked well in the geographic 
 coding of the census questionnaires; however, after preparation of 
 the ACG's were well underway (and the census date was too close 
 in time to permit change in the system), it was realized that there 
 were certain limitations to the file and some desirable improve- 
 ments should be made when feasible. The file, as initially 
 developed, would not permit comprehensive editing of the 
 records by computer. Also, because of the work involved in file 
 development and the time available, the ACG's X-Y coordinate 
 values and non-street information — railroad tracks, drainage 
 features, etc., reflected on the Metropolitan Map series and 
 capabilities of inclusion in computer mapping— were not planned 
 as part of the original ACG development but were scheduled for 
 post 1970 census improvement. 
 
 Using graph theory as the conceptual framework, the 
 solution combined the ACG file with information necessary to 
 describe the urban street network. By considering each street as a 
 
 line and each intersection as a node, the entire region could be 
 viewed as a series of interrelated nodes, lines, and enclosed areas. 
 Other features, such as streams or jurisdiction boundaries, could 
 also be defined in terms of segments and nodes. This approach 
 was called DIME (Dual Independent Map Encoding) and the 
 entire program is now referred to as the Geographic Base File 
 (DIME) System or GBF/DIME System. 
 
 The term DIME refers to the fact that the basic file is 
 created by coding two independent matrices (a) the nodes at the 
 end of the line segments and (b) the areas enclosed by the nodes 
 and line segments. 
 
 File Development 
 
 Before developing a GBF/DIME file, the Census Bureau's 
 Metropolitan Map series is reviewed and those places on the map 
 where a street or nonstreet feature intersects another feature, 
 comes to an end, or has a definite change of direction are 
 identified with a "node point" and labeled with a dot and a 
 unique identification number (fig. 1). A segment is that portion 
 of a street or non-street feature located between two nodes. It 
 includes the geographic and address codes pertaining to both sides 
 of the feature. Each segment can be described by its node points 
 as well as by its name. For example, nodes 71 and 77 identify a 
 section of Kimball Avenue. 
 
 A record is then prepared for each line segment which will 
 contain a series of geographic identifiers or code for the segment's 
 characteristics or for the area on either side of the segment. A 
 description of the geographic elements which appear in the file 
 follows this explanation. Figure 2 illustrates some of the 
 information contained in the file for the 100 block of Kimball 
 Avenue. 
 
 192 
 
GBF/DIME SYSTEM-THE DIFFUSION OF AN INNOVATION 
 
 Figure 1 
 
 ROSE 
 
 ^i l a 4 ron ,j l — - 
 
 /■IZER 
 
 -207 
 
 IIIG 
 
 20* 
 
 Figure 2 
 
 For each street segment a record contains: 
 
 From Node 
 
 71 
 
 To Node 
 
 77 
 
 Street Name 
 
 Kimball 
 
 Street Type 
 
 Avenue 
 
 Left Addresses 
 
 101-199 
 
 Right Addresses 
 
 100-198 
 
 Left Block 
 
 202 
 
 Left Tract 
 
 69 
 
 Right Block 
 
 203 
 
 Right Tract 
 
 69 
 
 By identifying each line segment, its node points, and the 
 enclosed area, this technique of geographic description facilitates 
 a computerized check for internal topological consistency. For 
 example, the computer can link the node points around a block 
 to determine whether all segments of the block are accounted for 
 
 and the codes assigned to each segment are complete and 
 consistent. 
 
 Although the file is now topological^ complete, it cannot 
 yet be translated into a map image. The x-y coordinates of each 
 node must be determined and entered on the corresponding 
 computer record to establish the spatial relation of the nodes to 
 the line segments (fig. 3). 
 
 Figure 3 
 
 LINE BOUNDARY 
 
 AREA 
 
 Use and Users 
 
 The GBF/DIME system is the result of 10 years of effort 
 and is the Bureau's prime geoprocessing resource. However, it has 
 become an important element for geographic processing for 
 agencies at the city, county, regional and State levels as well. 
 
 Correlating local physical, social, and economic data via 
 street addresses, geographic codes, and coordinates provides a 
 flexible framework for administration, not only within specific 
 agencies or programs, but also for analysis of relationships of 
 diverse variables, such as juvenile delinquency and school truancy. 
 Local data can also be compared to available Bureau data when 
 processed by the GBF/DIME method. 
 
 Some uses of the GBF/DIME system are— 
 
 A. In Urban Transportation Study programs the system is 
 used to locate by transportation zone the distribution 
 of motor vehicles that are registered within a com- 
 munity; the geographic origin and destination of daily 
 trips; and in the study of the journey to work the 
 distribution of employment which is used to study the 
 journey to work. 
 
 B. In criminal justice programs, local police departments 
 are using GBF/DIME to determine geographic patterns 
 of crime in order to optimize the design of police beats 
 and manpower allocation. The GBF/DIME program also 
 forms the basis for computerized dispatching systems of 
 police and fire departments. 
 
 C. In educational programs, local school boards use the 
 files to determine the geographic distribution of 
 children by grade to improve school facility planning 
 and revise school district boundaries in order to distri- 
 bute pupil loads more equitably. 
 
 D. In urban planning, many departments use the 
 GBF/DIME in their housing condition surveys to locate 
 deteriorating neighborhoods as a basis for planning 
 remedial action. 
 
 193 
 
INDEX 
 
 CENSUS, 1970 
 
 Currency of, 21 
 Magnitude of, 107 
 Problems 
 
 content, 115 
 
 coverage, 1 15 
 
 definitional, 112 
 
 geography, 109 
 
 interviewer recruitment, 108 
 Precedure for, 107, 109 
 Questionnaires 
 
 mail, 118 
 
 changes, 14 
 Release, levels of, 151 
 Response rate, 108 
 Synthetic estimates, 188 
 Undercount, 111 
 
 See also Survey Research, Current 
 Population Survey 
 
 CURRENT POPULATION SURVEY 
 
 Census, March "mini-", 1 17 
 
 Changes, 1 17 
 
 Decennial census, update to, 2, 13 
 
 Design of, 116 
 
 General, 116 
 
 Household sample, definition of, 6 
 
 Inflation, measures of, 49 
 
 Information collected, 118 
 
 Unemployment, details of, 117 
 
 DATA QUALITY 
 
 Accuracy checks, 1 1 5 
 
 Demographic analysis, 1 15 
 
 Error, 7 
 
 Imputation of data, 115 
 
 Interviewers 
 
 census, 1 18 
 
 private, 61 
 Questionnaires 
 
 mail, 83 
 
 general, 109 
 
 FEDERAL STATISTICAL SYSTEM 
 
 Committees, interagency, 143, 150 
 Components, 142 
 History, 138 
 Nature of, 138 
 Planning process 
 
 general, 142 
 
 future developments, 145 
 Problems with, 143 
 Programs, factors influencing, 143 
 
 See also Users, Reporting Burden, 
 Public Policy 
 
 GRAPHICS 
 
 Computerized mapping, 97 
 Data, accuracy, 95 
 Indicators, 89 
 Indices, 91 
 
 HEALTH STATISTICS 
 
 Commercial uses, 32 
 
 Collection," selective, 37 
 
 Currency, 39 
 
 Expert opinion, 166 
 
 Foreign, comparison with, 34 
 
 Goals for, 34 
 
 Geographic coding, 3 
 
 Law concerning, 30 
 
 Measurement, numerator and denominator, 1 
 
 Model, maternal infant care program, 180 
 
 National Center for, 30 
 
 Sources of, 31 
 
 Vital Statistics System, 31 
 
 See also Surveys, Reporting Burden, Users 
 
 HOUSING 
 
 Change, components of, 54 
 Familization status, 91 
 Measures over time, value of, 113 
 Moves, potential for tracking, 56 
 Neighborhood change, slight, 152 
 Trends, occupancy, 58 
 
 See also Small Areas, Census 1970, 
 Surveys 
 
 INNOVATIONS 
 
 GBF/DIME System, 189, 192 
 Perception of, 191 
 Stages of, 189 
 
 MARKETING 
 
 Advertising, direct mail, 151 
 Census data, relating names and 
 addresses to, 154 
 Goods, distribution of, 19, 154 
 
 See also Privacy 
 
 MINORITIES 
 
 Aliens, illegal, 111 
 Language problems, 108 
 Response rates, 7 
 Undercount, 111 
 
 NATIONAL ACCOUNTS 
 
 Basic concept, 41 
 
 Consumer Price Index, 146, 168 
 
 Criticisms, 42, 45 
 
 Detail of data produced for, 117 
 
 Economic measures, beyond, 42 
 
 Goals, purpose of, 41 
 
 Historic development, 41 
 
 Image of, 159 
 
 Measurement difficulties, 43 
 
 Schools of thought, 12 
 
 Uses, current, 46 
 
 See also Current Population Survey, 
 Social Indicators 
 
 PRIVACY 
 
 Aggregate information, distinction, 19 
 Census Bureau's mission, 25 
 
 Case histories, 27 
 Census, processing, 28 
 Census questionnaires 
 
 justification for detail, 26 
 
 release by Archives, 29 
 
 storage of, 29 
 Comparison, U.S. versus others, 18 
 Cost, 25, 60, 62 
 Current developments, 25 
 Definition of, 27 
 General, 8 
 Health records, 40 
 Legal aspects, 25 
 Legislative history, 26 
 Mailing lists, 19 
 
 removal from, 157 
 Manipulation, fears of, 23 
 Marketing 
 
 congressional concern, 152 
 
 techniques, 152 
 Names as census identifiers, 18, 109 
 National Center for Health Statistics, 30 
 Procedures for collected information, 61 
 Questions, intimacy may violate, 33 
 Record sharing, I RS, Social Security, 9 
 Statistical planning, effect on, 143 
 Technology can promote, 28 
 Telephone solicitors, 20 
 Universal identifier, Social Security, 29 
 
 PUBLIC POLICY 
 
 Census, congressional and administration 
 
 impact on, 112 
 Case study, Washington National Airport, 182 
 Federal Statistical System 
 
 interagency committees, 143, 150 
 budget, 143, 146 
 Fiscal restraints, 50 
 Privacy, philosophical problems, 17 
 Statistics, pressures on, 4 
 Timing of decisions, 186 
 Users, explanation of material, 59 
 
 REPORTING BURDEN 
 
 General, 142 
 
 Forms clearance, 149 
 
 Health surveys, 36 
 
 Need for information, 26 
 
 Research and development, 29 
 
 Study to determine, 147 
 
 SMALL AREAS 
 
 Area profile report system, 152 
 
 Census Bureau, 188 
 
 Change, 152, 154 
 
 Estimates of, 188 
 
 Graphic approach, applicable to, 89 
 
 Value of information, 158 
 
 194 
 
INDEX 
 
 SOCIAL INDICATORS 
 
 Cost of information, 1 
 
 Corporate involvement, criticisms of, 171 
 
 Corporate performance, measurement, 173 
 
 Corporation, socializing the, 170 
 
 Definitions of, 12, 104, 160 
 
 Goals 
 
 consensus on, 168 
 
 defining what to measure, 160, 179 
 
 techniques for discovery, 166 
 Health statistics, 39 
 History of, 162, 166 
 Images 
 
 reality of, 170 
 
 of society, 170 
 International concerns, 13, 105 
 Information 
 
 versus data, 151 
 
 tracking versus new data, 114 
 
 uses of, 54 
 Local areas, 169 
 
 Maximization, as a goal, 173, 179 
 Measurement 
 
 case study, 167 
 
 available data, uses of, 174 
 
 economic welfare, 45 
 Numerator and denominator as 
 measures, 1, 154 
 Rank order 
 
 indices, 54 
 
 key factors, 57 
 Social Area Analysis, 90 
 Unemployment, 168 
 
 See also Graphics, Health Statistics, National 
 Accounts, Social Observatories 
 
 SURVEY RESEARCH 
 
 Advantages of, 77 
 
 Analytic strengths, 81 
 
 Case studies, 81 
 
 Data linkage, primary and secondary, 81 
 
 Ecological fallacy, 81 
 
 Firms, selection of, 86 
 
 Sample selection, 5, 80 
 
 SURVEY TECHNIQUES 
 
 Analysis, logic of, 78 
 
 Delphi, polling the experts, 166 
 
 Echo, 79 
 
 Likert, 79 
 
 Null hypothesis, 87 
 
 Randomized response, 33 
 
 Scaling, 79 
 
 Semantic differential, 79 
 
 SURVEYS 
 
 Census comparison with R.L. Polk, 53 
 Consumer expenditure, 26 
 Criminal victimization, 119 
 Health 
 
 patient examination, 31 
 
 physicians, 34 
 
 international, 110 
 Household income, 57 
 Population, annual estimates, 188 
 Postcensal, 112 
 
 See also Census 1970, Current 
 Population Survey 
 
 USERS 
 
 Age Search, 1 10 
 
 Access to data, 148 
 
 Census Bureau orientation toward, 14 
 
 Census questions orientation toward, 155, 158 
 
 Census Use Study, 59 
 
 Corporate performance, measurement 
 
 data available, 174 
 Data dissemination, 39, 147 
 Federal Statistical System, advisory 
 
 committees, 145 
 Geographic coding, 2, 14 
 Health statistics, 35, 38 
 Information, changing needs, 156, 158 
 Society, innudated with data, obligation 
 
 to potential users, 4 
 State and local cooperation, 148 
 Statistical definitions, inputs to, 5 
 Statistics, using those available 
 
 efficiently, 21 
 
 See also Public Policy 
 
 SOCIAL OBSERVATORIES 
 
 Census Bureau, 13 
 Definitional problems, 15 
 Need for, 1 1 
 Social indices, link with, 13 
 
 See also Social Indicators 
 
 SYSTEMS ANALYSIS 
 
 Firms, selection of, 186 
 General, 177 
 
 Models, criticisms of, 179 
 Transfer function, 177 
 
 195 
 


 
technical papers 
 
 No. 1. The Sample Survey of Retail Stores: A Report on Methodology 
 
 ... Description of the Sample for the Monthly Retail Trade Report (unnumbered revision) 
 Accuracy of Census Statistics With and Without Sampling 
 Farm Population, 1880-1950 
 The Post Enumeration Survey: 1950 
 
 Tests and Revisions of Bureau of the Census Methods of Seasonal Adjustments 
 The Current Population Survey Reinterview Program, Some Notes and Discussion 
 The Current Population Survey, A Report on Methodology 
 Trends in the Income of Families and Persons in the United States: 1947-1960 
 Reconciliation of the 1958 Census of Retail Trade with the Monthly Retail Trade Report 
 Population Trends in the United States: 1900 to 1960 
 
 Response Errors in Collection of Expenditures Data by Household Interviews: An Experimental Study 
 Estimating Trading-Day Variation in Monthly Economic Time Series 
 Sampling Application in Censuses of Population and Housing 
 
 The International Standard Industrial Classification and the U.S. Standard Industrial Classification 
 The X-11 Variant of the Census Method II Seasonal Adjustment Program 
 Present Value of Estimated Lifetime Earnings 
 
 Trends in the Income of Families and Persons in the United States: 1947 to 1964 
 Changes Between the 1950 and 1960 Occupation and Industry Classification 
 The Current Population Survey Reinterview Program, January 1961 through December 1966 
 Correlation Between United States and International Standard Industrial Classifications 
 Characteristics of America's Engineers and Scientists: 1960 and 1962 
 
 Measures of Overlap of Income Distributions of White and Negro Families in the United States 
 The Position of United States in World Commodity Exports in 1968 
 The Annual Survey of Manufactures: A Report on Methodology 
 Demographic Computer Library 
 
 1970 Occupation and Industry Classification System Showing Sources 
 Sampling Applications of the 1970 Census Publication, Maps, and Public Use Summary Files 
 From the Old to the New 1972 SIC for Establishments 
 
 Scientific and Technological Development Activities of the Bureau of the Census 
 Census County Division, Past and Future 
 
 Consistency of Reporting of Ethnic Origin in the Current Population Survey 
 Standards for Discussion and Presentation of Errors in Data 
 Characteristics of Persons in Engineering and Scientific Occupations: 1972 
 Indexes to Survey Methodology Literature 
 
 Family (Money) Income 1947 to 1971 : Summarizing Twenty-Five Years of a Summary Statistic 
 Census Bureau Programs To Measure Consumer Purchase Expectations: 1959-1973 
 The Census Bureau, A Numerator and Denominator for Measuring Change 
 Comparison of Persons of Spanish Surname and Persons of Spanish Origin in the United States 
 
 To receive further information or to order copies of these reports, contact the nearest district office 
 of the Department of Commerce or write to - 
 
 SUBSCRIBER SERVICES SECTION (PUBLICATIONS) 
 SOCIAL AND ECONOMIC STATISTICS ADMINISTRATION 
 WASHINGTON, D.C. 20233 
 
 No. 
 
 2 
 
 No. 
 
 3 
 
 No. 
 
 4 
 
 No. 
 
 5 
 
 No. 
 
 6 
 
 No. 
 
 7. 
 
 No. 
 
 8 
 
 No. 
 
 9 
 
 No. 
 
 10 
 
 No. 
 
 11. 
 
 No. 
 
 12, 
 
 No. 
 
 13. 
 
 No. 
 
 14. 
 
 No. 
 
 15. 
 
 No. 
 
 16. 
 
 No. 
 
 17. 
 
 No. 
 
 18. 
 
 No. 
 
 19. 
 
 No. 
 
 20. 
 
 No. 
 
 21. 
 
 No. 
 
 22. 
 
 No. 
 
 23. 
 
 No. 
 
 24. 
 
 No. 
 
 25. 
 
 No. 
 
 26. 
 
 No. 
 
 27. 
 
 No. 
 
 28. 
 
 No. 
 
 29. 
 
 No. 
 
 30. 
 
 No. 
 
 31. 
 
 No. 
 
 32. 
 
 No. 
 
 33. 
 
 No. 
 
 34. 
 
 No. 
 
 35. 
 
 No. 
 
 36. 
 
 No. 
 
 37. 
 
 No. 
 
 38. 
 
 J 
 
U.S. DEPARTMENT OF COMMERCE 
 
 Social and Economic Statistics Administration 
 
 BUREAU OF THE CENSUS 
 Washington. DC. 20233 
 
 r 
 
 PE 7ii s i 1 i A i}r. u , N ir 1 f.^ ,Tv '-.Bs; 
 
 US DE 
 
 OFFICIAL BUSINESS 
 
 A000070^3fl 
 
 J 
 
 working papers 
 
 Raw Materials in the United States Economy: 1900-52 (Preliminary) 
 
 The Role of the 1954 Census of Manufactures in Overcoming Problems of Industry Data 
 
 Tract Data Compared for a 25-Percent Sample and a Complete Census 
 
 Sampling in the 1950 Census of Population and Housing 
 
 Occupational Trends in the United States: 1900 to 1950 
 
 Raw Materials in the United States Economy 
 
 Papers Presented at the Census Tract Conference, December 29, 1958 
 
 Materials on the Preparation and Conduct of the U.S.S.R. All-Union Population Census of 1959 
 
 Historical Comparability of Census of Manufactures Industries: 1929- 1958 
 
 Papers Presented at the Census Tract Conference, December 29, 1959 
 
 Papers Presented at the Census Tract Conference, August 25, 1960 
 
 Papers Presented at the Census Tract Conference, August 28, 1961 
 
 Papers Presented at the Census Tract Conference, September 8, 1962 
 
 The Spectral Analysis of Economic Time Series 
 
 Methodology and Scores of Socioeconomic Status 
 
 Procedural Report on the 1960 Censuses of Population and Housing 
 
 Papers Presented at the Census Tract Conference, September 5, 1963 
 
 The Measurement of Performance Potential in Manufacturing Establishments 
 
 Tests of Use of Post Office Resources to Improve Coverage of Censuses 
 
 Industry Classification and Sector Measures of Industrial Production 
 
 A Spectral Study of "Overadjustment" for Seasonality 
 
 Papers Presented at the Census Tract Conference, December 29, 1964 
 
 Spectral Analysis and Parametric Methods for Seasonal Adjustment of Economic Time Series 
 
 Self-Enumeration as a Method for the 1970 Census of Housing 
 
 Measuring the Quality of Housing 
 
 Changes in the Structure of Manufacturing Employment 
 
 Methodology of Consumer Expenditures Survey 
 
 Metropolitan Area Definition 
 
 Survey Applications of Social Psychological Questions 
 
 Raw Materials in the United States Economy: 1900 1966 
 
 Price Variation in New FHA Houses: 1959-1961 
 
 Pretests and Dress Rehearsals of the 1970 Census of Population and Housing: A Procedural 
 
 History 
 
 An Estimate of a Quasi-Stable Age-Sex Distribution for Ghana in 1960 
 
 Investigation of Census Bureau Interviewer Characteristics, Performance, and Attitudes: A 
 
 Summary 
 
 Raw Materials in the United States Economy: 1900-1969 
 
 Response Variance in the Current Population Survey 
 
 Who's Home When 
 
 Economic Censuses of the United States: Historical Development 
 
 Population and Housing Inquiries in U.S. Decennial Censuses, 1790-1970 
 
 Studies in Occupational and Industrial Classification 
 
 To receive further information or to order copies of these reports, contact the nearest district office 
 of the Department of Commerce or write to - 
 
 SUBSCRIBER SERVICES SECTION (PUBLICATIONS) 
 SOCIAL AND ECONOMIC STATISTICS ADMINISTRATION 
 WASHINGTON, D.C. 20233 
 
 No. 
 
 1 
 
 No. 
 
 2 
 
 No. 
 
 3 
 
 No. 
 
 4 
 
 No. 
 
 5 
 
 No. 
 
 6 
 
 No. 
 
 7. 
 
 No. 
 
 8 
 
 No. 
 
 9. 
 
 No. 
 
 10. 
 
 No. 
 
 11. 
 
 No. 
 
 12 
 
 No. 
 
 13 
 
 No. 
 
 14 
 
 No. 
 
 15 
 
 No. 
 
 16. 
 
 No. 
 
 17 
 
 No. 
 
 18. 
 
 No. 
 
 19 
 
 No. 
 
 20 
 
 No. 
 
 21 
 
 No. 
 
 22. 
 
 No. 
 
 23. 
 
 No. 
 
 24 
 
 No. 
 
 25. 
 
 No. 
 
 26 
 
 No. 
 
 27 
 
 No. 
 
 28 
 
 No. 
 
 29 
 
 No. 
 
 30 
 
 No. 
 
 31. 
 
 No. 
 
 32 
 
 No. 
 
 33 
 
 No. 
 
 34 
 
 No. 
 
 35 
 
 No. 
 
 36 
 
 No. 
 
 37 
 
 No. 
 
 38 
 
 No. 
 
 39 
 
 No. 
 
 40