y/ technical paper 41 AN EVALUATION OF 1970 CENSUS OCCUPATIONAL CLASSIFICATION The Postcensal Manpower Survey — Census Match Study U.S. Department of Commerce « , BUREAU OF THE CENSUS \ technical paper 41 AN EVALUATION OF 1970 CENSUS OCCUPATIONAL CLASSIFICATION The Postcensal Manpower Survey — Census Match Study Issued February 1978 U.S. Department of Commerce Juanita M. Kreps, Secretary BUREAU OF THE CENSUS Manuel D. Plotkin, Director BUREAU OF THE CENSUS Manuel D. Plotkin, Director Robert L. Hagan, Deputy Director Daniel B. Levine, Associate Director for Demographic Fields POPULATION DIVISION Meyer Zitter, Chief ACKNOWLEDGMENTS The PMS-Census Match Study was conducted in the Labor Force Statistics Branch, Paula J. Schneider, Chief. The principal participants were Thomas J. Palumbo and Victor M. Valdisera, who jointly authored this report. Mary Kay Friday developed the computer program to select the sample. David Bateman and William Smith, both of the Statistical Methods Division, provided technical assistance in the calculation of various statistical measures. Overall direction was provided by Murray S. Weitzman, Assistant Division Chief, Socioeconomic Statistics Programs. Library of Congress Cataloging in Publication Data United States. Bureau of the Census. An evaluation of 1970 census occupational classification. (Technical paper— Bureau of the Census; 41 ) 1. Occupations— Classification. 2. United States— Occupations. 3. Scientists— United States. I. Title. II. Series: United States. Bureau of the Census. Technical paper; 41 . HB2595.U52 1977a 331.1'1'0973 77-26803 For sale by Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402, or any U. S. Department of Commerce district office. Postage stamps not acceptable; currency submitted at sender's risk. Remittances from foreign countries must be by international money order or by a draft on a U.S. bank. Stock Number 003-024-01485-8 Contents Page Introduction 1 Summary of observations 1 Background of the study 2 Correspondence between the census and the PMS occupational categories 3 The sample 4 Reasons for disagreement 6 The response 6 The classification system: collection vehicles 6 The classification system: classification schemes 6 Definitions of mismatch causes 7 I. Classification-system causes 8 II. Response causes 9 Summary of the causes of mismatches 11 Analysis of results 11 All occupations 12 Operations and computer specialists 14 Engineers 15 Mathematical specialists 15 Life scientists . 17 Physical scientists 17 Social scientists , 19 Graphic analysis 19 TEXT TABLES Table A. All occupational groups by causes of mismatches between the census and PMS occupational classifications, by level of mismatch 13 B. Operations and computer specialists by causes of mismatches between census and PMS occupational classifications, by level of mismatch 14 C. Engineers by causes of mismatches between census and PMS occupational classifications, by level of mismatch 16 D. Mathematical specialists by causes of mismatches between census and PMS occupational classifications, by level of mismatch 16 E. Life scientists by causes of mismatches between census and PMS occupational classifications, by level of mismatch 18 F. Physical scientists by causes of mismatches between census and PMS occupational classifications, by level of mismatch 18 G. Social scientists by causes of mismatches between census and PMS occupational classifications, by level of mismatch 20 III Contents — Continued DETAILED TABLES Table Page 1. Detailed occupation in 1970 according to the 1970 census, by detailed occupation in 1970 according to the 1972 Postcensal Manpower Survey, for respondents in the 1972 Postcensal Manpower Survey 24 2. Total cases within universe of the study, by agreement between census and PMSoccupational classification, by level of disagreement for mismatched cases, by detailed census occupation in 1970 27 3. Universe and sample cases by major 1970 census occupational group, by level of mismatches 28 ILLUSTRATIONS Illustration A. Census occupational items 3 Illustration B. PMS occupational items 4 Illustration C. Occupational titles for chemists 7 Illustration D. Schematic representation of the C-scale 12 EXAMPLES Example 1 . Correspondence between census and PMS occupational categories 5 Example 2. Causes of mismatches 8 CHARTS Figure 1. Major 1970 census occupational groups, by C-scale categories 21 Figure 2. Correctly categorized mismatches (according to the C-scale), by major 1970 census occupational groups, by causes of mismatches 22 Figure 3. Incorrectly categorized mismatches (according to the C-scale), by major 1970 census occupational groups, by causes of mismatches 23 APPENDIXES A. Postcensal Manpower Survey (PMS) occupational coding scheme: "List C— Occupations" 29 B. Decision-logic table for comparing PMS and census responses 31 C. Census rules for coding occupation 33 D. Influence of professional identification on reporting of PMS occupation 36 E. Reconciliation process 38 F. Sample design, estimation procedure, and reliability of the estimates 41 IV An Evaluation of 1970 Census Occupational Classification The Postcensal Manpower Survey — Census Match Study INTRODUCTION The Bureau of the Census is continually involved in efforts to evaluate and improve its occupational classification system The 1972 Postcensal Manpower Survey (PMS), conducted by the Bureau under the sponsorship of the National Science Founda- tion, enabled the Bureau to make a significant contribution to this ongoing effort. This report presents the methodology and substantive results of a study, known as the PMS-Census Match, that attempted to take advantage of this opportunity. The sample for the 1972 PMS was chosen from among persons who had been identified in the 1970 Census of Population as being in selected engineering, scientific, and technical occupations in the 1970 experienced civilian labor force (ECLF). Each person had been clerically coded to a detailed 1970 census occupational category on the basis of replies to a set of questions about "current or most recent job activity." In the 1972 survey, these persons were asked to respond to a similar set of questions about their 1970 job. Each person, however, was also requested to select the most appropriate detailed occupational category for this job. In this way, two classifications of the respondent according to occu- pation in 1970 were made available. It was possible to compare the person's census and PMS categories, and to use the comparison to judge whether the person had been correctly classified in the census. A person's PMS classification, of course, either agreed or disagreed with the census classification. When it agreed, in- creased confidence could be placed in the belief that the census had coded the person into an occupational category that accurately reflected the kind of work he or she was doing. When, however, it disagreed, a doubt arose about the validity of the census classification. This doubt would remain until the reason for the difference could be established. It was assumed that one possible cause of disagreement could arise when the person's occupation was described in the PMS differently from the way it was described in response to the census questions. Further research on such cases would reveal whether the census response or the PMS response had provided a more accurate identification of the person's occupation in 1970, or whether the descriptions were in conflict to such a degree that the more appropriate identification could not be determined. Also, it could be that the census and PMS descriptions were essentially the same, but differences in the way the descriptions were collected or categorized had led to the occupational classi- fication differences. SUMMARY OF OBSERVATIONS Certain strengths and weaknesses of the census occupational classification system were revealed when it was placed against the background of the PMS system. Additionally, placing the census against the PMS procedure was tantamount to con- trasting two different methods of classifying occupations: the census method, in which clerical coders translated responses into occupational categories; and the PMS method, in which the classification was done by the respondents themselves. This contrast was especially important for its insights into the pitfalls of the PMS approach and the ways such pitfalls might be avoided. In particular, the following observations, concerned with specific means of improving the Census and the PMS occu- pational classification systems, were made during the study. Their meaning will become clear to the reader once further sections of this report are read. The first six observations are related to the census system; the last three, to the PMS system. Observation 1. It should be stressed to the census coders that information in all sections of the census occupational question should be considered before they assign an occupa- tional code. Frequently, the coders placed a respondent in an occupation based solely on the written entry in question 34a (See illustration A); this practice led to misclassification errors, especially when the respondent supplied insufficient informa- tion in question 34a. Thus, any proposal to reinstate the "cascade rule," (see page 9) or any variation of it, should not be approved. Observation 2. An extensive examination should be con- ducted concerning the use of the lowest-code rule (see page 9) in assigning occupational codes to respondents whose jobs involve both managerial activities and activities related to specific occupations such as electrical engineering. Because of the lowest-code rule, the respondents were arbitrarily assigned to particular scientific or engineering occupations, rather than to managerial occupations. A better approach may be the use of the job title section (question 34c of illustration A) as the deciding factor. A respondent who enters a job title that is consistent with the managerial activities reported in question 34b should be classified into a managerial occupation. Otherwise, the person should be coded to a specific scientific or engineering occupation. Observation 3. Extensive research needs to be done on the major occupational group, "operations and system researchers and analysts," since a significant number of persons are improperly classified into this group. The major cause of the problem is the failure of respondents to differentiate among the various kinds of systems analysts, such as "business" systems analyst or "computer" systems analyst. One partial solution may be the use of industry requirements. For example, a respondent could be classified as a computer systems analyst if the written entry (or entries) is "systems analyst" and the industry code is either 189 ("manufacturing, electronic com- puting equipment") or 739 ("computer programming services"). Another possible solution is to classify "systems analyst, not specified" under the occupation "computer systems analyst"; of course, doing so could mean that some persons who legitimately belong to the group "operations and systems researchers and analysts" would be misclassified. Observation 4. A further examination should be made of certain occupational titles that were identified as problem areas. The following ^re some examples: a. There does not seem to be any real difference between the title "mathematical actuary" (now included under "036 mathematicians") and the title "actuary" (now included under "035 actuaries"). b. "Psychiatric social worker" should be an occupational title under "psychologists" rather than under "social workers." Many psychiatric social workers coded them- selves to the occupation "psychologists" on the PMS. Also, many reported on the PMS that their major field of study was psychology. c. Some of the occupational titles, such as "wildlife biologist," "fishery biologist," and "plant pathologist," could be moved from the occupation "agricultural scientists" to "biological scientists." This change may make the titles under both of these census occupations more homogeneous. d. The job title "financial analyst" might be placed more appropriately under the occupation "accountants" than under "economists." The PMS indicates that most financial analysts identified themselves as "accountants" rather than as "economists." Furthermore, most of these persons reported in the PMS that their major field of study was accounting. Observation 5. The respondent who reports a dual occu- pation in every part of question 34 should be placed into a residual category rather than into either of the specific occupations. For example, the entry "programmer-analyst" should be coded to "005 computer specialists, not elsewhere classified," instead of to "003 computer programmers" or to "004 computer systems analysts." An occupational title "programmer-analyst" should be added to the list of titles for the occupation "computer specialists, not elsewhere classified." Observation 6. Some occupational titles should be added to the census classification scheme. "Marketing representative," "microscope operator," and "behavioral science teacher" are additions that should be made. Observation 7. Most PMS coding errors involve college and university teachers of engineering and science who failed to code themselves to their specific fields of engineering or science as instructed on PMS List C (See appendix A). Instead, these persons placed themselves into the occupational category "451 teachers, college and university, excluding engineering and science." A possible solution would be to underscore the instruction "including college professors and instructors" which is stated on List C after each of the major engineering and scientific occupational groups. Another suggested solution is to change the present code 451 category to read "Nonscience and nonengineering college and university teachers (Engineering and science teachers, see codes 401-432 above)." Observation 8. A substantial proportion of the mismatches occur because of structural differences between the census and PMS classification schemes. The limited size of the PMS occupational coding list is probably the major reason. As explained in a later section, the PMS classification scheme did not provide the respondents with the occupational titles, such as "financial analyst," included under the List C categories, whereas the census did provide these titles to the census coders. Thus, respondents often misclassified their occupational titles in the PMS One possible way to minimize these PMS misclassi- fications would be to add some examples of the appropriate occupational titles to each of the PMS occupational categories. For example, the occupational titles, "botanist," "entomologist," "bacteriologist," could be listed next to the PMS occupational category "biological scientists." This listing would be especially helpful to respondents who are trying to determine whether they belong in one of the residual categories of List C, such as "other social scientists.'' Observation 9. Many respondents incorrectly used PMS residual categories (codes 412, 416, 428, and 432) when their occupational titles were included under specific PMS occu- pational categories. It is hoped that the suggestion made in observation 8 will prevent some of these misclassifications. It may, however, be advisable during the processing stage of future PMS surveys to verify clerically a sample of cases in which the respondents use residual categories. This verification would give the analyst some data concerning the reliability of the counts in these residual groups. BACKGROUND OF THE STUDY The official title of the 1972 Postcensal Manpower Survey (PMS) is the "1972 Professional, Technical, and Scientific Manpower Survey." The survey was conducted by the Bureau of the Census during the spring and summer of 1972. 1 The sample for the survey was chosen from among persons enumerated on either a 15- or 5-percent sample questionnaire in the 1970 census. 2 The sample included approximately 97,000 persons who had been classified by the 1 970 census as being in the 1 970 experienced civilian labor force (ECLF) in one of 64 target occupations. This study is restricted to those members of the sample who were in one of 44 engineering and science occupations and who reported their 1970 occupation in the 1972 PMS; there were approximately 35,000 such persons. The occupational classifications in the 1970 census are based upon responses to items 33, 34, and 35 on the 1970 census questionnaire. These questions are reproduced in illustration A. 1 For detailed information on the survey, see U.S. Bureau of the Census, Characteristics of Persons in Engineering and Scientific Occupations: 1972, Technical Paper No. 33, U.S. Government Printing Office, Washington, D.C. 1974. ? See U.S. Bureau of the Census, U.S. Census of Population & Housing: 1970, Procedural History, PHC(R)-11, Washington, D.C. 1976, Chapter 15, for a description of the various questionnaires used in the 1970 census. Persons were classified according to the system described in the publication, 1970 Census of Population Alphabetical Index of Industries and Occupations. 3 Parts of this system essential to an understanding of this study are explained in the following sections of this report. The 1970 PMS occupational classifications are based on responses to item 22 on the PMS questionnaire. This item, along with related PMS items, is shown in illustration B. Item 22 asked the respondent to specify the kind of work being done in each of his or her three most recent jobs, beginning with the job held in 1972 (or nearest to 1972, if the respondent was not working in 1972) and working backwards. The person's occu- pation during the time period comparable to that of the 1970 census was selected from this job history. Respondents answered item 22 by entering a code and a description from the reference list (List C, reproduced in appendix A) that accom- panied the PMS questionnaire. The essential features of the PMS occupational classification system are also explained below. Table 1 presents a distribution of the 1970 census science or engineering occupations of the PMS respondents in this study, by their detailed occupations in 1970 according to the 1972 PMS. Had the PMS and the census classified persons into corresponding categories, all cases in the table would be within cells located on the diagonal. This study was undertaken to learn why significant numbers of cases are located in off- diagonal cells. This research does not permit statements about the overall accuracy of the census figures on the number of persons in each of these science or engineering occupations. For a particular census occupational category, the study was concerned only with the cases for which the PMS classification was in disagreement (i.e., the off-diagonal cases). Since the PMS indicated that these cases did not properly belong in the census category, the cases were reviewed and the most appropriate classification of PMS and census responses was determined. Other census occupational groups, however, were not viewed for evidence, based on the PMS, that some of their members should have been classified to the census category being examined. Nor were the on-diagonal cases reviewed for evidence of misclassi- fication in either source. The study, then, was one-sided; and the PMS estimate of how many persons belong in a census category is within the scope of this research, bounded on its upper limit by the number in the particular census occupation and on its lower limit by the agreement cases. CORRESPONDENCE BETWEEN THE CENSUS AND THE PMS OCCUPATIONAL CATEGORIES The first stage of the research was to establish the corre- spondence between the census and the PMS occupational categories. The PMS categories equivalent to each census category were determined; this correspondence or equivalence was established at both the major-group and detailed levels of occupational classification. The census detailed occupational category "economists," for example, was considered to be 3 The full citation is U.S. Bureau of the Census, 1970 Census of Population Alphabetical Index of Industries and Occupations, U.S. Government Printing Office, Washington, D.C., 1971. Illustration A 33-35. Current or most recent job activity Describe clearly this person's chief 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 I960. 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 fan 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? 35. (For example: Types, keeps account books, pies, sells cars, operates printing press, cleans buildings, finishes concrete) What was his job title? 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 Local government employee (city, enmity, etc.) ... O Self employed in own business, professional practice, or farm — Own business not incorporated Own business incorporated Working without pay in family business or farm equivalent to the detailed PMS occupational category "econo- mists"; the major census group "social scientists" corresponded to the major PMS group "social scientists." The correspondence between the two classification systems is shown in example 1 and by the diagonal of table 1. Persons whose occupations Illustration B Pari IV - EMPLOYMENT PROFILE and working back. Please include all jobs, no! just scientific or significant chongn in your duties, level of responsibility, or occ for all three jobs. If you had more than one regular job last week second current regular job should be reported in question 37 ) our last THREE civilian jobs beginning with the major job you held last week (or the last job you held) technical jobs. In answering these questions, consider a change in |obs to have occurred if there were upation, even though you may have rontinued working for the same employer. Please answer each question , report on the one which you considered to be your primary or most important job. (Information about 20 For whom d.d you work' Name of company, business, organization, government LAST CIVILIAN JOB Job held last week or most recent job Job A SECOND-TO-LAST CIVILIAN JOB Job B THIRD-TO-LAST CIVILIAN JOB Job C Name Name Name City or county 1111 1 1 City or cale Structural Differences Managerial-Concept Methodological Differences PMS Insufficient Responses Conflicting Responses Census Insufficient Responses 1 Census Coding Errors Other Methodological Differences Note: In this schematic, the C-Scale is shown as a number line whose values range from +1 to —1 . The absolute magnitudes of the numbers signify degrees of certainty and the signs of the numbers signify correctness. Thus, all values fall between an absolute value of "0", indicating complete uncertainty, and an absolute value of "1 ", indicating complete certainty. The "+" sign attached to a value indicates that the census categorization is correct; and the "— " sign, that it is incorrect. 13 two-thirds are caused by census coding errors, and the remaining third by census insufficient responses or by other meth- odological differences. An incorrect census categorization at the major-group level, of course, represents a much more serious failing of the census classification system than does one at the detailed level. In view of the initial indications, as shown in tables 1 and 2, the discovery that only 29 percent of the major-group level mismatches are incorrectly classified is encouraging. This 29 percent represents only about 16 percent of all mismatches (major-group and detailed levels). In about 80 percent of the mismatches, therefore, the census classification scheme succeeded in placing persons at least within the correct major occupational group. At the detailed level of mismatches, table A reveals that as many as 37 percent of the cases may be correctly classified in the census, with almost all of these cases involving structural differences, managerial-concept methodological differences, or PMS coding and processing errors. The table shows that 38 percent of the mismatches are incorrectly categorized in the census, 24 percent because of census coding errors, 12 percent because of census insufficient responses, and about 3 percent because of other methodological differences. Census coding error appears to be a significant cause of census misclassification, constituting about two-thirds of the cases for which the evidence from the sample indicates that the census system has failed. Errors, of course, occur in all coding operations, some stemming from systematic causes and others occurring through chance oversights. It could not be established conclusively into which of these two categories any particular census coding error fell. Nevertheless, this research indicates that a major cause of coding errors is the failure, for systematic reasons, of the coding clerks to use all the information available to them. Three such systematic reasons were identified. First, coders often placed a person in an occupational category based solely on the written entry in the first part of the three-part occupational question, even if this entry was insufficient. Second, coders tended to key on one word of the written response when determining the occupational category of the respondent; for example, a coder would spot the word "statistical" in question 34a and assign the person to the detailed occupational category "statisticians" instead of using all the available information in question 34a to 34c, which would have caused the respondent to be coded to "statistical clerks." Third, when coding some occupational titles, coders did not consult the Alphabetical Index because they incorrectly assumed that they knew the occupational category of the title in question; for example, the occupational title "computer programming manager" was often coded to "computer pro- grammers" instead of to its proper occupational category, "computer systems analysts." A second definition of "correctness" is mentioned above, and it is suggested that structural and methodological differ- ences may be indexes of how well the 1 970 census classification system succeeded according to this definition. Structural dif- ferences, it is said, may indicate disagreement over the classifi- cation of various occupational titles or characteristics. In this regard, subcategory misclassifications, a type of structural Table A. All Occupational Groups by Causes of Mismatches Between the Census and PMS Occupational Classifications, by Level of Mismatch (For meaning of symbols, see text) Causes of mismatch Mismatched cases Standard error Major-group level mismatches Standard error Detailed-level mismatches Standard error All causes, total. U.262 100.0 Errors, total Census coding errors. PMS coding errors 2 . . . Response causes, total Conflicting responses Census insufficient responses. PMS insufficient responses. . . . Structural differences, total, PMS subcategory misclassifications PMS residual -category classifications Methodological differences. Managerial concept Other 393 31.1 273 21.6 120 9.5 46 5 36.8 319 25.3 121 9.6 25 2.0 304 24.1 165 13.1 139 11.0 96 7.6 75 5.9 21 1.7 (X) 1.3 1.2 0.8 1.4 1.2 0.8 0.4 1.2 0.9 0.9 0.7 0.7 0.4 l 696 249 140 109 248 174 56 18 113 71 42 82 75 7 100.0 35.8 20.1 15.7 35.6 25.0 8.0 2.6 16.2 10.2 6.0 11.8 10.8 1.0 (X) 1 .8 1.5 1.4 1 .8 1.6 1.0 0.6 1.4 1.1 0.9 1.2 1.2 0.4 566 144 133 11 217 145 65 7 191 94 97 14 (X) 14 100.0 25.4 23.5 1.9 38.3 25.6 11.5 1.2 33.7 16.6 17.1 2.5 (X) 2.5 (X) 1.8 1.8 0.6 2.0 1.8 1.3 0.4 2.0 1.6 1.6 0.7 (X) 0.7 'Includes 4 cases incorrectly included in the sample. Includes processing errors. 14 difference would seem to indicate a more serious disagreement than residual-category classifications. In subcategory misclassifi- cations, persons indicated that their occupations were in a different category from the one that had been assigned in the census, whereas in residual-category classifications they indi- cated simply that the census category was not appropriate without specifying what category would have been more appropriate. The results show that structural differences are about equally divided between the two subcategories, with subcategory misclassifications comprising about 13 percent of all mismatches and residual-category classifications about 11 percent. The entire group of structral differences constitute about 24 percent of all mismatches; this fact could mean that in as many as a quarter of the mismatches the census failed to re- flect, at the detailed level of occupational classification, the kind of work the persons were doing. As expected, structural differences occur more frequently at the detailed level than at the major-group level. It would appear to be more likely that persons would agree with the census that a particular title or set of occupational characteristics should be classified in the major group "engineers," for example, than that it should be placed in the detailed occupation "chemical engineers" rather than in "mining and petroleum engineers." In fact, structural differences do account for a larger proportion of cases among the detailed-level mismatches (34 percent) than among the major-group level ones (16 percent). However, about 6 percent of the major-group level mismatches are residual- category classifications; this 6 percent is particularly significant because, for all of these cases, the PMS respondents indicated that their work was entirely outside the fields of engineering or science. These persons entered codes 436 ("other health occupations"), 448 ("technicians, other fields") or 469 ("other occupations, not specified above"). Methodological differences comprise about 8 percent of all mismatches; most of these differences are managerial-concept methodological differences (6 percent of all mismatches). By definition, a managerial-concept mismatch exists only at the major-group level; at this level, table A shows that this cause is responsible for about 11 percent of such mismatches. Other methodological differences cause about 2 percent of all mismatches. OPERATIONS AND COMPUTER SPECIALISTS Among operations and computer specialists, response problems and classification-system causes each contribute approximately half to the mismatch universe (table B). About 46 percent of all mismatches involve either conflicting or insufficient responses. Also, nearly one-third of all mismatches in this group contain errors, primarily census coding errors. Structural and meth- odological differences together cause approximately 22 percent of the classification differences. Viewing these results in terms of the C-scale, about half of the operations and computer specialist mismatches are mis- classified in the census at either the major-group or detailed levels. In other words, mismatches arising from census coding errors, census insufficient responses, or other methodological differences are assigned to incorrect census occupational categories. Table B. Operations and Computer Specialists by Causes of Mismatches Between Census and PMS Occupational Classifications, by Level of Mismatch (For meaning of symbols, see text) Causes of mismatch Mismatched cases Standard error of percent Major-group level mismatches Standard error of percent Detailed- level mismatches Standard error of percent All causes, total Errors, total Census coding errors PMS coding errors 1 ... Response causes, total Conflicting responses Census insufficient responses... PMS insufficient responses Structural differences, total PMS subcategory misclassifications PMS residual-category classifications Methodological differences, total. Managerial concept Other 1 Includes processing errors. 263 100.0 82 31.2 75 28.5 7 2.7 23 46.8 72 27.4 49 18.6 2 0.8 37 14.1 23 8.7 14 5.3 21 8.0 12 4.6 9 3.4 (X) 2.9 2.8 0.9 3.1 2.7 2.4 0. i 2.1 1.7 1.4 1.6 1.3 1.1 ] 12 100.0 30 26.8 25 22.3 5 4.5 49 43.8 25 22.3 24 21.4 21 18.8 14 12.5 7 6.3 12 10.7 12 10.7 (X) 4.2 1.9 4.7 3.9 3.9 3.7 3.1 2.3 2.9 2.9 151 100.0 52 34.4 50 33.1 2 1.3 74 49.0 47 31.1 25 16.6 2 1.3 16 10.6 9 6.0 7 4.6 9 6.0 (X) (X) 9 6.0 (X) 3.9 3.8 0.9 4.1 3.8 3.0 0.9 2.5 1.9 1.7 1.9 (X) 1.9 15 One of the most common census occupational misclassifi- cations occurred when persons reported their occupation in the census to be "systems analyst," but did not indicate any type of specialty, e.g., "computer systems analyst," "business systems analyst." These persons were assigned to the census occupa- tional category "operations and systems researchers and analysts." The information on their PMS form, however, indicates that the proper occupation for the vast majority is "computer systems analysts." A partial solution for this par- ticular kind of census insufficient response may be the use of industry requirements in the coding process. For example, a respondent could be classified as a "computer systems analyst" if the written entry is "systems analyst" and the industry code is 189 ("manufacturing, electronic computing equipment") or 739 ("computer programming services"). Another possible solution is to add the occupational title, "systems analyst, n.s.," to the occupational category, "computer systems analysts." Of course, this addition could mean that some persons who legitimately belong to the category "operations and systems researchers and analysts" would be misclassified. Another common census occupational misclassification occurred when the coders incorrectly assigned the written entry "computer programming manager" to the occupational category "computer programmers" instead of to the category "computer systems analysts." Although coding instructions in the 1970 census specified that "manager" was a keyword in assigning correct occupation codes, this rule should receive more emphasis in the future. At the other end of the C-scale are those mismatches for which the census occupational code assignments are correct. For mismatches involving structural differences (PMS subcategory misclassifications and PMS residual-category classifications), PMS coding errors, PMS insufficient responses, and managerial- concept methodological differences, the assumption of correct census occupational categorizations can be made. These cases comprise about 22 percent of the operations and computer specialist mismatches. Finally, about 27 percent of the operations and computer specialists mismatches are caused by conflicting responses. There is no way to judge, using the PMS information, whether the occupational category assigned in the census is correct for these cases. When the sample is divided into major-group level and detailed-level mismatches, there are some differences in their respective C-scale patterns. Major-group level mismatches are more likely to be correctly categorized in the census than are detailed-level mismatches. In fact, only 13 percent of the mismatches at the detailed level are correctly categorized in the census, compared with 34 percent of the cases at the major- group level. ENGINEERS About 60 percent of the mismatches for engineers stem from classification-system causes (errors, structural differences, and methodological differences) and about 40 percent from response problems (table C). The classification-system causes are divided between errors— mostly census coding errors— and systemic differences. About two-thirds of the response causes involve conflicting responses; the remaining one-third involve census insufficient responses. PMS insufficient responses are a very minor problem for the engineering group. In terms of the C-scale, about 36 percent of all engineer mismatches are correctly classified in the census. About two-thirds of these correctly categorized cases involve structural differences. At the major-group level, managerial-concept meth- odological differences and structural differences are responsible for similar proportions of the correctly classified cases. Interestingly, the subcategory, managerial-concept meth- odological differences, contains a higher proportion of mismatches at the major-group level for engineers than it does for any other occupational group. At the other end of the C-scale, about 38 percent of all engineer mismatches were placed into an incorrect occupational category in the census, with census coding errors being the major reason for census misclassification. In most of the cases containing census coding errors, the respondent described one specific engineering occupation but was coded to another specific engineering occupation. For example, in one case the respondent provided the written description "mechanical engineer" in census questions 34a-c, but was given the code for the occupation "civil engineers." This problem illustrates that non-systematic coding error is a major problem in some census occupation groups. Only a stringent quality control system can minimize such errors. Another type of census occupational misclassification in- volves respondents who report their occupations to be "design engineer" on the census. In these cases a specific engineering occupation, such as "civil engineers," was assigned according to the industry reported in questions 33 a-c. In other words, two persons who supplied the same occupational description, "design engineer," could have been coded to different detailed categories if they reported different industries on the census. Although this study did not determine how many persons are correctly classified using this criteria, there seems to be evidence that a sufficient number of persons are misclassified to warrant further research in this area. Finally, conflicting responses are responsible for about 27 percent of the mismatches in the total sample. Conflicting responses are present in this same proportion (about 27 percent) at the major-group and detailed levels. A common conflicting response is that of persons who reported "sales engineer" in the census, and "salesman" or a specific engineering occupation, such as, "mechanical engineer," in the PMS. As is stated in an earlier section, the PMS information is of little use in determining the accuracy of the census occupational categori- zations for these cases. MATHEMATICAL SPECIALISTS Classification-system causes account for about 64 percent 9 of all mismatch cases for the mathematical specialists, whereas response problems are present in only about 36 percent 9 of the cases (table D). Errors are the major type of classification-system 'These percentages are based on 140 cases. The 144 cases shown in table D contain four cases incorrectly included among the mismatches. 16 Table C. Engineers by Causes of Mismatches Between Census and PMS Occupational Classifications, by Level of Mismatch (For meaning of symbols, see text) Causes of mismatch Mismatched cases Standard error of percent Major-group level mismatches Standard error of percent Detailed- level mismatches Standard error of percent All causes, total. 275 100.0 Errors , total. Census coding errors. PMS coding errors 1 ... Response causes, total Conflicting responses Census insufficient responses. PMS insufficient responses. . . . Structural differences, total. PMS subcategory misclassif ications PMS residual-category classifications Methodological differences, total. Managerial concept Other 76 27.6 69 25.1 7 2.5 106 38.5 73 26.5 32 11.6 1 0.4 65 23.6 39 14.2 26 9.5 28 10.2 25 9.1 3 1.1 (X) 2.7 2.6 0.9 2.9 2.6 1.9 0.4 2.5 2.1 1.8 1.8 1.7 0.5 112 100.0 30 26.8 23 20.5 7 6.3 34 30.4 29 25.9 5 4.5 2 17.9 7 6.3 13 11.6 28 25.0 25 22.3 3 2.7 (X) 4.3 3.8 2. 3 4.3 4. 1 2.0 3.6 2.3 3.0 4. 1 3.9 1.5 L63 100.0 46 28.2 46 28.2 72 44.2 44 27.0 27 16.6 1 0.6 45 27.6 32 19.6 13 8.0 (X) (X) (X) 3.5 3.5 3.9 3.5 2.4 0.6 3.5 3.1 2.1 (X) 1 Includes processing errors. Table D. Mathematical Specialists by Causes of Mismatches Between Census and PMS Occupational Classifications, by Level of Mismatch (For meaning of symbols, see text) Causes of mismatch Mismatched cases Standard error of percent Major-group level mismatches Standard error of percent Detailed-level mismatches 2 Standard error of percent All causes, total. l 144 100.0 Errors, total Census coding errors. PMS coding errors 3 ... Response causes, total Conflicting responses Census insufficient responses. PMS insufficient responses. . . . Structural differences, total. PMS subcategory misclassif ications PMS residual-category classifications Methodological differences, total. Managerial concept Other 7 5 52.1 4 1 28.5 34 23.6 50 34.7 44 30.6 3 3.5 1 0.7 5 3.5 1 0.7 4 2.8 10 6.9 6 4.2 4 2.8 (X) 4.2 3.8 3.5 4.0 3.8 1.5 0.7 1.5 0.7 1.4 2.1 1.7 1.4 l 115 100.0 63 54.8 )0 26.1 33 28.7 38 33.0 32 27.8 5 4.3 1 0.9 4 3.5 1 0.9 3 2.6 6 5.2 6 5.2 (X) 4.6 4.1 4.2 4.4 4.2 L.9 0.9 1.7 0.8 1.5 2.0 2.0 2 29 100.0 12 41.4 11 37.9 1 3.4 12 41.4 12 41.4 4 (X) 4 1.4 3.4 13.8 (X) 13.8 (X) (X) (X) (X) (X) (X) (X) (X) (X) (X) (X) (X) (X) (X) 1 Includes 4 cases incorrectly chosen for the sample, not shown separately. 2 The sample of detailed-level mismatches for mathematical specialists includes the total number of cases (29) in the universe; thus, the figures in this category are not subject to sampling errors, and standard errors do not apply. 3 Includes processing errors. 17 cause, with census coding errors being responsible for a slightly larger proportion of errors than are PMS coding errors. The other kinds of classification causes, structural and meth- odological differences, are only minor reasons for misclassifi- cation. Conflicting responses are the most frequent type of response problem (about 9 out of every 10 response problem cases contain conflicting responses). Interpreting these data in light of the C-scale reveals that it is likely that about 32 percent of the mathematical specialists mismatched cases are correctly classified in the census. The majority of the correctly classified cases contain PMS coding errors. One of the most common PMS coding errors occurred at the major-group level when college or university professors of mathematics classified themselves to "college or university teachers, excluding science or engineering," instead of to "mathematicians," "statisticians," or "actuaries." A similar proportion (about 35 percent) of mismatches are incorrectly classified in the census. Census coding errors are at the base of most of the misclassifications. One of the more widespread census coding errors involved clerical workers, such as mathematical clerks, who were classified as "mathematical specialists." Misclassifications of mathematical specialists caused by insufficient census responses are often closer to being correct than are the same kind of misclassifications of other occupa- tional groups. The insufficient responses for the mathematical specialists frequently contain the words "statistical" or "mathematical." Such words usually narrow the possible occupational categorizations to a choice between a particular professional mathematical specialist occupation (such as "statistician") and the occupation "statistical clerk." Insuf- ficient responses of other occupational groups often permit a greater number of choices among possible categorizations. In the middle of the C-scale are the conflicting responses (about 30 percent of all mismatches.) Proportionately, these cases are more frequent at the detailed level than at the major-group level. Although a limited amount can be said about the correctness of conflicting response cases, there is one suggestion that may alleviate one kind of conflicting response problem. The title "mathematical actuary," which is included under "mathematicians" in the 1970 Classified Index, perhaps should be moved to the occupation "actuaries." There seems from this investigation to be no real difference between the title "mathematical actuary" and the title "actuary." Finally, two important facts distinguish the mathematical specialists from other occupational groups. First, standard errors are not applicable at the detailed level since all cases in the universe were examined. Second, PMS List C does not provide a residual category for the mathematical specialists. Thus, struc- tural differences are a very minor problem for this occupation group. LIFE SCIENTISTS For life scientists (table E) most of the mismatches are the result of classification-system causes (about 70 percent) rather than the result of response problems (30 percent). Among the various categories of classification-system causes, errors (about 31 percent) and structural differences (about 35 percent) account for similar proportions of the total mismatched cases. About two-thirds of the response problems for this group are con- flicting responses; only about one-third are census or PMS insufficient responses. Analyzing the data from table E in terms of the C-scale shows that about 56 percent of the mismatches are placed in correct census occupations. Structural differences are the most im- portant element in the correctly classified cases, representing about 60 percent of these cases. A common structural difference concerned persons who reported their PMS occupational title to be "wildlife biologist" or "fishery biologist." In these cases, the respondents would code themselves on the PMS to "biological scientists," even though their occupation titles were subcategories of "agri- cultural scientists" according to the census. Although the census correctly classified these people according to its 1970 occu- pational scheme, a question does arise whether it placed persons who reported certain occupational titles, such as "wildlife biologist," into an occupational category that best reflects the kind of work they were doing. One possible way to solve this problem is to switch some of the occupational titles, such as "wildlife biologist," "fishery biologist," and "plant pathol- ogist", from "agricultural scientists" to "biological scientists." These changes may make the titles under both of these occupations more homogeneous. It is not known, however, what effect this change would have on the match cases. 1 ° At the other end of the C-scale are mismatches for which the census occupational code assignments are incorrect. For the mismatches involving census coding errors, census insufficient responses, or other methodological differences, there is little doubt thatthey are misclassified in the census. These cases make up about 24 percent of the mismatches for life scientists. One of the most frequent census coding errors for this group involves persons who reported their occupation to be either "wildlife biologist" or "fishery biologist" on the census, and were given the code for "biological scientists" instead of the one for "agricultural scientists." It is not possible to determine why the coders made this error. The coders may have assumed that these titles belonged to the occupation "biological scientists," or the errors may have been caused by chance oversights. Finally, about 20 percent of the life scientists mismatches arise from conflicting responses. Conflicting responses were present in the same proportions (about 20 percent) at the major-group and detailed levels. PHYSICAL SCIENTISTS Classification-system causes (about 64 percent) are more preva- lent than response problems (about 36 percent) for the total sample of physical scientists (table F). Of the three types of classification-system causes, structural differences are the most common, followed by errors, and then by methodological differences. When the sample is divided into major-group and detailed-level cases, their respective distributions by the causes 10 That is, cases in which the respondent reported one of the occupational titles such as "wildlife biologist," on the PMS and entered the code for "agricultural scientists." 18 Table E. Life Scientists by Causes of Mismatches Between Census and PMS Occupational Classifications, by Level of Mismatch (For meaning of symbols, see text) Causes of mismatch Mismatched cases Standard error of percent Major-group level mismatches Standard error of percent Detailed-level mismatches Standard error of percent All causes, total , Errors , total , Census coding errors. PMS coding errors 1 , Response causes, total Conflicting responses , Census insufficient responses. . , PMS insufficient responses Structural differences, total..... PMS subcategory misclassif ications PMS residual-category classifications Methodological differences, total, Managerial concept Other 1 Includes processing errors. 185 100.0 57 30.8 33 17.8 24 13.0 53 28.6 36 19.5 11 5.9 6 3.2 64 34.6 18 9.7 46 24.9 11 5.9 10 5.4 1 0.5 (X) 3.4 2.8 2.5 3.3 2.9 1.7 1.3 3.5 2.2 ).2 1.7 L.6 0.4 100.0 42 36.8 21 18.4 21 18.4 36 31.6 24 21.1 6 5.3 6 5.3 25 21.9 18 15.8 7 6.1 11 9.6 10 8.8 1 0.9 (X) 4.5 3.6 3.6 4. 3 3.8 2.1 2. 1 3.9 3.4 2.2 2.8 2.6 0.8 71 100.0 15 21.1 12 16.9 3 4.2 17 23.9 12 16.9 5 7.0 39 (X) 54.9 54.9 (X) (X) 4.8 4.4 2.4 5.1 4.4 3.0 5.9 5.9 (X) Table F. Physical Scientists by Causes of Mismatches Between Census and PMS Occupational Classifications, by Level of Mismatch (For meaning of symbols, see text) Causes of mismatch Mismatched cases Standard error of percent Major-group level mismatches Standard error of percent Detailed-level mismatches Standard error of percent All causes, total Errors, total. Census coding errors PMS coding errors 1 Response causes, total Conflicting responses Census insufficient responses. . , PMS insufficient responses Structural differences, total..... PMS subcategory misclassif ications PMS residual -category classifications Methodological differences, total. Managerial concept. Other , 'includes processing errors. 211 100.0 49 23.2 25 11.8 24 11.4 76 36.0 65 30.8 8 3.8 3 1.4 73 34.6 35 16.6 38 18.0 13 6.2 11 5.2 2 0.9 (X) 2.9 2.2 2.2 3.3 3.2 1. 3 0.8 3.3 2.5 2.6 1.6 1.5 0.5 12 1 100.0 38 31.4 18 14.9 20 16.5 54 44.6 45 37.2 7 5.8 2 1.7 16 13.2 8 6.6 6.6 10.7 9. 1 1.7 (X) 4.2 3.2 3.4 4.5 4.4 2.1 1.2 3.1 2.2 2.2 2.8 2.6 1. 1 100.0 11 7 4 12.2 7.8 4.4 22 20 1 1 24.4 22.2 1. 1 1. 1 57 63.3 27 30.0 30 33.3 (X) (X) (X) 3.4 2.8 2. 1 4.5 4.4 1.1 1.1 5.1 4.8 5.0 (X) 19 of mismatch differ in some ways from the one for the total sample. For example, at the detailed level proportionally more cases contain classification-system causes than do cases in the sample as a whole. Furthermore, structural differences are much more frequent at the detailed level than at the major-group level, whereas clerical errors are more widespread at the major-group level than at the detailed level. In terms of the C-scale, it is likely that about one-half (about 53 percent) of the mismatches are assigned correct census occupations. The largest group of these correctly classified cases involves structural differences (PMS residual-category classifi- cations and PMS subcategory misclassification). The following are examples of the most common cases with structural differences: (1) persons who reported their PMS occupation to be "astronomers" but placed themselves in the PMS residual category, "other natural scientists" instead of the proper category, "physicists"; (2) respondents who stated on the PMS that they were "meteorologists" and incorrectly coded them- selves to "earth and marine scientists" instead of to "other natural scientists". The other end of the C-scale shows that only about 1 7 percent of the mismatches are misclassified in the census. The remaining 31 percent of the cases have conflicting census and PMS occupations, which prevents any assignment of census correctness. A frequent conflicting response concerns persons who reported the occupational title "metallurgist" on the census and "metallurgical engineer" on the PMS. SOCIAL SCIENTISTS Among the entire sample of social scientists (table G), classification-system causes are responsible for about 70 percent of the mismatches and response problems for about 30 percent. Although this same basic split is present at the major-group and detailed levels, there are differences among the various kinds of classification-system causes at each of these levels. For instance, structural differences occur more often at the detailed level (about 53 percent) than at the major-group level (about 22 percent), whereas errors are a greater problem at the major- group level (about 38 percent) than at the detailed level (about 13 percent). In terms of the C-scale, it is fairly certain that about 58 percent of the social scientists mismatches are placed in correct census occupations. As was true for most of the other occupational groups, structural differences are the largest component of the correctly classified cases. The two most typical cases involving structural differences are: (1) persons who reported "financial analyst" in the PMS and coded themselves to "accountants" rather than to "economists"; (2) persons who described their PMS occupational title as "psychiatric social worker" and placed themselves in the occupation "other social scientist" instead of in "other occu- pations." Although in both these examples these persons are correctly classified in the census according to the 1970 classification system, there is considerable evidence from the PMS that the occupational title "financial analyst" should be moved from the occupational category "economists" to that of "accountants." Also, there is some support from the PMS for placing the occupational title "psychiatric social worker" under "psychologists" rather than under "social workers." Both of these changes may reflect more realistic occupational categories for these occupation titles. In terms of the C-scale, about 26 percent of the mismatches are incorrectly classified in the census, with census coding errors causing the largest proportion of the misclassifications. A common census misclassification concerns persons who reported their occupational title to be "marketing representative" and were incorrectly coded to the occupational category "econo- mists." It should be mentioned that the occupational title, "marketing representative" does not appear in either the 1970 Alphabetical or Classified Indexes. Finally, about 16 percent of the mismatches contain con- flicting responses. An interesting form of conflicting response occurs among a few respondents who identified themselves as being "psychologists" in the PMS but "physiologists" in the census. Almost certainly their intended census entry of the word "psychologist" was misspelled as "physiologists." Most likely, they were enumerated in one of the field followups conducted during the census operations, and for some reason, their occupational information was incorrectly recorded by clerical personnel. GRAPHIC ANALYSIS Figures 1, 2, and 3 summarize the results of the PMS-Census Match in the form of bar charts. Figure 1 distributes the mismatches in each occupational group according to the categories of the C-scale. Figure 2 distributes the correctly categorized areas of figure 1 according to mismatch causes, and figure 3 does the same thing for the incorrectly categorized areas of figure 1. 20 Table G. Social Scientists by Causes of Mismatches Between Census and PMS Occupational Classifications, by Level of Mismatch (For meaning of symbols, see text) Causes of mismatch Mismatched cases Standard error of percent Major-group level mismatches Standard error of percent Detailed- level mismatches Standard error of percent All causes , total Errors, total Census coding errors PMS coding errors 1 Response causes , total Conflicting responses Census insufficient responses. . PMS insufficient responses Structural differences, total.... PMS subcategory misclassif ications PMS residual-category classifications Methodological differences, total Managerial concept Other 1 Includes processing errors. I 84 100.0 54 29.3 30 16.3 24 13.0 57 31.0 29 15.8 16 8.7 12 6.5 60 32.6 49 26.6 11 6.0 13 7.1 11 6.0 2 1.1 (X) 3.3 2.7 2.5 3.4 2.7 2. 1 1.8 J. 4 3.3 1.7 1.9 1.7 0. 7 122 100.0 46 37.7 23 18.9 23 18.9 37 30.3 19 15.6 9 7.4 9 7.4 27 22.1 23 18.9 4 3.3 12 9.8 11 9.0 1 0.8 (X) 4.4 3. 3 3.5 4.2 3.3 2.4 2.4 3.8 3.5 L.6 2.7 2.6 0.8 62 100.0 8 12.9 7 11.3 1 1.6 20 32.3 10 16.1 7 11.3 3 4.8 33 53.2 26 41.9 7 11.3 1 1.6 (X) (X) 1 1.6 (X) 4.3 4.0 1.6 5.9 4.7 4.0 2.7 6.3 6.3 4.0 1.6 (X) 1.6 Figure 1 Major 1970 Census Occupational Groups, by C-scale Categories Percent 100 Social Scientists Correctly Categorized Conflicting Responses ^ Not Applicable 21 Figure 2 Percent Correctly Categorized Mismatches (According to the C-scale), by Major 1970 Census Occupational Groups, by Causes of Mismatches Physical Scientists Social Scientists Managerial Concept Methodological Differences PMS Coding Errors ^ PMS Insufficient Responses 22 Figure 3 Percent 100 Incorrectly Categorized Mismatches (According to the C-scale), by Major 1970 Census Occupational Groups, by Causes of Mismatches Social Scientists Census Census Other Coding Insufficient Methodological Errors Responses Differences 23 24 Table 1. DETAILED OCCUPATION IN 1970 ACCORDING TO THE 1970 CENSUS, BY DETAILED OCCUPA- TION IN 1970 ACCORDING TO THE 1972 POSTCENSAL MANPOWER SURVEY, FOR RESPONDENTS IN THE 1972 POSTCENSAL MANPOWER SURVEY Detailed 1970 census occupatlc Detailed Postcensal Manpower Survey (PMS) occupati With 1970 PMS xupati pute: pro- C'Hll- puter sys- tems lysts Com- puter puter spe- search lysts and archi- tec- tural Total Operations and computer specialists Computer programmers Computer systems analysts Computer specialists, n.e.c Operations and systems research- ers and analysts Engineers Aeronautical and astronautical engineers Chemical engineers Civil engineers Electrical and electronic engineers Industrial engineers Mechanical engineers Metallurgical and materials engineers Mining engineers Petroleum engineers Sales engineers Engineers, n.e.c Engineering teachers* Mathematical specialists Actuaries Mathematicians Statisticians Mathematics teachers* Life scientists Agricultural scientists Agriculture teachers* Biological scientists Biology teachers* Foresters and conservationists 1 . . Physical scientists Atmospheric and space scientists. Atmospheric, earth, marine, and space teachers* Geologists Marine scientists Chemists Chemistry teachers* Physicists and astronomers Physics teachers* Life and physical scientists, n.e.c Social scientists Economists Economics teachers* Psychologists Psychology teachers* Sociologists Sociology teachers* Political scientists Urban and regional planners Social scientists, n.e.c Social science teachers, n.e.c*. n.e.c. Not elsewhere classified. "College and university. 'With 4 or more years of college 6,960 2,123 2,172 467 2,198 18,606 2,154 2,048 2,333 2,549 2,207 2,259 590 178 377 1,933 1,789 2,178 189 275 784 930 3,215 476 173 1,100 794 672 5,104 226 185 769 134 2,039 302 850 548 51 4,824 1,843 298 991 457 53 257 23 lib 118 448 34,938 5,979 1,854 1,865 398 1,862 15,909 1,852 1,822 2,014 2,177 1,882 1,925 516 144 338 1,595 1,475 169 1,868 170 237 663 798 2,775 400 147 952 673 603 4,392 198 158 645 114 1,795 250 729 460 43 4,015 1,522 256 814 377 41 217 18 289 100 181 4,182 3,774 1,654 1,458 238 424 196 1,256 1,027 160 1,684 362 943 737 234 320 107 10,716 837 1,413 1,317 659 1,463 23 1,591 22 1,050 12 1,440 43 400 2 85 - 261 - 633 9 942 58 121 5 ,773 149 1,051 25 Table 1. DETAILED OCCUPATION IN 1970 ACCORDING TO THE 1970 CENSUS, BY DETAILED OCCUPA- TION IN 1970 ACCORDING TO THE 1972 POSTCENSAL MANPOWER SURVEY, FOR RESPONDENTS IN THE 1972 POSTCENSAL MANPOWER SURVEY— Continued Detailed 1970 Detailed Postcensal Manpower Survey (PMS) 1970- -Continued With 1970 PMS occupation reported — Continued Metal- lur- gical and Mining and petro- Life and physical Total Operations and computer specialis Computer programmers Computer systems analysts Computer specialists, n.e.c... Operations and systems research ers and analysts Engineers Aeronautical and astronautical engineers Chemicals engineers Civil engineers Electrical and electronic engineers Industrial engineers Mechanical engineers Metallurgical and materials Mining engineers Petroleum engineers Sales engineers Engineers, n.e.c Engineering teachers"* Mathematical specialists Actuaries Mathematicians Statisticians Mathematics teachers* Life scientists Agricultural scientists......... Agriculture teachers"" Biological scientists Biology teachers* , Foresters and conservationists 1 , Physical scientists , Atmospheric and space scientist; Atmospheric, earth, marine, and space teachers"" Geologists Marine scientists , Chemists , Chemistry teachers"' , Physicists and astronomers Physics teachers*. Life and physical scientists, Social scientists , Economists , Economics teachers* Psychologists Psychology teachers* Sociologists Sociology teachers* Political scientists Urban and regional planners Social scientists, n.e.c Social science teachers, n.e.c." 276 112 5 2 2,197 266 187 217 300 253 341 120 104 346 lib 132 117 1,724 207 529 68 1,123 177 465 306 7 74 J 145 .".■■I ] 17 26 Table 1. DETAILED OCCUPATION IN 1970 ACCORDING TO THE 1970 CENSUS, BY DETAILED OCCUPA- TION IN 1970 ACCORDING TO THE 1972 POSTCENSAL MANPOWER SURVEY, FOR RESPONDENTS IN THE 1972 POSTCENSAL MANPOWER SURVEY— Continued Detailed 1970 census occupation Detailed Postcensal Manpower Survey (PMS) occupati 1970— Contin With 1970 PMS occupation reported — Conti Life and physical scientists — Continued Other life Psy- cholo- gists Soci- olo- gists and an- thro- polo- 1970 PMS occupati not reported I. n- ployed labor force Total Operations and computer specialist: Computer programmers , Computer systems analysts Computer specialists, n.e.c , Operations and systems research- ers and analysts , Engineers , Aeronautical and astronautical engineers , Chemical engineers , Civil engineers Electrical and electronic engineers Industrial engineers , Mechanical engineers Metallurgical and materials engineers Mining engineers Petroleum engineers Sales engineers Engineers, n.e.c Engineering teachers* Mathematical specialists Actuaries Mathematicians Statisticians Mathematics teachers* Life scientists Agricultural scientists Agriculture teachers* Biological scientists Biology teachers " Foresters and conservationists 1 ., Physical scientists Atmospheric and space scientists. Atmospheric, earth, marine, and space teachers* Geologists Marine scientists Chemists Chemistry teachers ' Physicists and astronomers Physics teachers* Life and physical scientists, n.e.c Social scientists Economists . Economics teachers* Psychologists Psychology teachers * Sociologists Sociology teachers ' Political scientists Urban and regional planners Social scientists, n.e.c Social science teachers, n.e.c*. 103 Ha 1,546 267 117 624 189 26 119 366 247 1 11 579 1 79 945 4h0 1,039 120 107 20 2,208 1,074 134 172 178 12 95 5,949 ,697 981 144 269 306 307 102 69 375 302 423 226 528 319 521 372 775 325 462 334 338 314 225 121 312 132 980 440 178 76 59 26 395 148 212 121 136 69 sin 321 »a 27 Table 2. TOTAL CASES WITHIN UNIVERSE OF THE STUDY, BY AGREEMENT BETWEEN CENSUS AND PMS OCCUPATIONAL CLASSIFICATION, BY LEVEL OF DISAGREEMENT FOR MIS- MATCHED CASES, BY DETAILED CENSUS OCCUPATION IN 1970 Detailed 1970 census occupation Number Percent Number Percent Mismatched cases Number Percent Major-group level Number Percent Detailed level Number Percent Total cases in universe Operations and computer specialists Computer programmers. Computer systems analysts Computer specialists, n.e.c...... Operations and systems researchers and analysts Engineers Aeronautical and astronautical engineers. Chemical engineers Civil engineers Electrical and electronic engineers. Industrial engineers. Mechanical engineers Metallurgical and materials engineers Mining engineers Petroleum engineers Sales engineers Engineers , n.e.c ...„„. Engineering teachers* ............ Mathematical specialists Actuaries. Mathematicians Statisticians Mathematics teachers* Life scientists Agriculture scientists. .......... Agriculture teachers* Biological scientists. ........... Biology teachers* Foresters and conservationists 1 . . Physical scientists Atmospheric and space scientists. Atmospheric, earth, marine, and space teachers*. ............ Geologists Marine scientists. ............... Chemists. ........................ Chemistry teachers*. ............. Physicists and astronomers....... Physics teachers* Life and physical scientists, n.e.c............... Social scientists Economists Economics teachers*. ............. Psychologists. ................... Psychology teachers*. ............ Sociologists Sociology teachers* Political scientists Urban and regional planners Social scientists, n.e.c......... Social science teachers, n.e.c*. 34,938 5,979 1,854 1,865 398 15,909 ,852 ,822 ,014 ,177 ,882 ,925 516 144 338 ,595 ,475 169 1,868 170 237 663 798 2,775 400 147 952 673 603 4,392 198 158 645 114 1,795 250 729 460 43 4,015 1,522 256 814 377 41 217 18 289 100 381 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 15,318 2,182 1,027 943 134 78 715 872 1,121 1,201 533 990 224 67 191 341 289 121 957 117 84 340 416 1,438 145 73 368 405 44 7 2,629 38 108 515 55 1,046 171 410 285 1,447 247 113 579 179 22 111 6 39 30 121 43.8 36.5 55.4 50.6 33.7 4.2 41.9 38.6 47.9 55.7 55. 28, 51, 43, 46, 56. 21.4 19.6 71.6 51.2 68.8 35.4 51.3 52.1 38.7 60.2 74.1 59.9 19.2 68.4 79.8 48.2 58.3 68.4 56.2 62.0 2.3 36.0 16.2 44.1 71.1 47.5 53.7 51.1 33.3 13.5 30.0 31.8 19,620 3,797 827 922 264 9,244 1,137 950 89 3 976 1,349 935 292 77 147 1,254 1,186 48 911 53 153 323 382 1,337 255 74 584 268 156 1,763 160 50 130 59 749 79 319 175 42 2,568 1,275 143 235 198 19 106 12 250 70 260 56.2 63.5 44.6 49.4 66.3 95.8 58.1 61.4 52.1 44.3 44.8 71.7 48.6 56.6 53.5 43.5 78.6 80.4 28.4 48.8 31.2 64.6 48.7 47.9 48.2 63.8 50.3 61.3 39.8 25.9 40.1 31.6 20.2 51.8 41.7 31.6 43.8 38.0 97.7 64.0 83.8 55.9 28.9 52.5 46.3 48.8 66.7 86.5 70.0 68.2 13,187 2,096 197 382 157 1,360 383 495 514 569 828 429 113 58 76 951 485 48 882 50 133 317 382 1,216 218 71 521 253 153 1,575 99 50 118 48 708 77 278 158 39 2,469 1,255 139 190 188 15 98 12 248 68 256 37.7 35.1 10.6 20.5 39.4 73.0 31.1 20.7 27.2 25.5 26.1 44.0 22.3 21.9 40.3 22.5 59.6 32.9 28.4 47.2 29.4 56.1 47.8 47.9 43.8 54.5 48.3 54.7 37.6 25.3 35.9 50.0 31.6 18.3 42.1 39.4 30.8 38.1 34.3 90.7 61.5 82.5 54, 23, 49, 36, 45. 66, 85.8 68.0 67.2 6,433 1,701 630 540 107 424 4,295 754 455 379 407 521 r )06 179 19 71 303 701 121 37 3 63 15 3 n.e.c. Not elsewhere classified. *College and university. x With 4 or more years of college. 28 TABLE 3. UNIVERSE AND SAMPLE CASES BY MAJOR 1970 CENSUS OCCUPATIONAL GROUP, BY LEVEL OF MISMATCH Mismatched cases Major 1970 census occupation group Major-group level Detailed -group level Total Sample Total Sample 13,187 2,096 4,949 882 1,216 1,575 2,469 696 112 112 115 114 121 122 6,433 1,701 4,295 29 121 188 99 571 : 154 l 164 29 71 90 l 63 'These numbers differ from those in tables A,B,C, and G because they include cases whose PMS or census questionnaires could not be located . APPENDIX A Postcensal Manpower Survey (PMS) Occupational Coding Scheme List C— Occupations" ill ■ ListC - OCCUPATIONS This list is to be used in answering the questions about the kind of work you were doing and about your professional or occupational classification. When the instructions for a particular item on the questionnaire request you to enter a code and description from this list, please scan the entire list, then choose the appropriate entry. If you cannot find exactly the right entry, please choose the one that comes nearest to it. If none of the entries is at all appropriate, use the "Other" category (code 469) and enter a brief description in the space provided on the questionnaire. Codt Description Engineers, including college professors and instructors 401 Engineer, aeronautical and astronautical 402 Engineer, agricultural 403 Engineer, chemical 404 Engineer, civil and architectural 405 Engineer, electrical and electronic 406 Engineer, industrial 407 Engineer, mechanical 408 Engineer, metallurgical and materials 409 Engineer, mining and petroleum 410 Engineer, nuclear 411 Engineer, enviromental and sanitary 412 Engineer, other fields (Describe briefly under the applicable item on the questionnaire.) Computer Specialists, including college professors and instructors 413 Computer programmer 414 Computer systems analyst 415 Computer scientist 416 Other computer specialist (Describe briefly under the applicable item on the questionnaire.) Mathematicians ond Statisticians, including collegf professors and instructors 417 Actuary 418 Mathematician 419 Statistician 420 Operations research analyst Natural Scientists, including college professors and instructors 421 Agricultural scientist, including foresters and conservationists 422 Biological scientist 423 Biochemist 424 Chemist 425 Earth and marine scientist, including geoglogists, geophysicists, oceanographers, etc. 426 Medical scientist, excluding persons who are primarily medical practitioners; see Health Occupations below 427 Physicist 428 Other natural scientist (Describe briefly under the applicable item on the questionnaire.) Social Scientists, including college professors and instructors 429 Economist 430 Psychologist 431 Sociologist or anthropologist 432 Other social scientist (Describe briefly under the applicable item on the questionnaire.) Code Description Health Occupations, including persons who are primarily practitioners. Persons engaged primarily in medical research, teaching, and similar activities use code 426, Medical scientist. 433 Physician or surgeon 434 Technician, dental 435 Technician, medical 436 Other health occupation (Describe briefly under the applicable item on the questionnaire.) Technicians and Technologists, except medical 437 Designer, electronic parts and machine tools 438 Designer, industrial 439 Designer, other 440 Draftsman 441 Surveyor 442 Technician, biological and agricultural 443 Technician, electrical and electronic 444 Technician, construction, highways, and architectural 445 Technician, mechanical 446 Technician, other engineering 447 Technician, physical science 448 Technician, other fields (Describe briefly under the applicable item on the questionnaire.) Teachers 449 Teacher, elementary school 450 Teacher, secondary school 451 Teacher, college and university, excluding engineering and science 'Engineering and science teachers see codes 401-432 above.) Administrators, Managers, and Officials, excluding farm 452 College president or dean 453 Administrator or manager, scientific and technical research and development 454 Administrator or manager, production and operations 455 Administrator, manager, or official, all other, excluding sel f-employ ed 456 Self-employed proprietor All Other Occupations 457 Accountant 458 Attorney or judge 459 Clerical or sales worker (such as salesman, bookkeeper, secretary, etc.) 460 Clergyman 461 Craftsman (such as baker, carpenter, electrician, mechanic, repairman, etc.) 462 Farmer (owner, manager, tenant, or farm laborer) 463 Fireman or policeman 464 Laborer, except farm 465 Librarian 466 Merchant or shopkeeper, self-employed 467 Operative (such as assembler, factory worker, miner, welder, truck driver, etc.) 468 Postal worker 469 Other occupations, not specified above (Describe briefly under the applicable item on the questionna ire.) FORM PMS-3 (10-26-71) GP 9 25-537 30 APPENDIX B Decision-Logic Table for Comparing Postcensal Manpower Survey (PMS) and Census Responses A response in the Postcensal Manpower Survey consists of (1) written entries in questions 22, 23, and 24, and (2) a code from reference List C (see appendix A) in the code box of question 22. The PMS code, however, is also part of the PMS classifi- cation system; and often, it is difficult to determine the extent to which the respondent considered it to be a part of the occupational description, rather than merely a way of classi- fying this description. It is for this reason that a decision-logic table was created to help translate the PMS response into one that could be compared with the census response. The decision- logic table provided a consistent, scientific method of deciding the weight that should be given to the PMS code as an element of the occupational description; it specified under what con- ditions the code would be considered entirely as a classification device, and under what conditions it would be considered an integral, and perhaps deciding, piece of occupational description information There are two sections in the decision-logic table. The first, consisting of situations 1-7, refers to cases whose written entries on the PMS convey essentially the same information as the census entries. The second section consisting of situations 8-14, refers to the converse cases— those whose PMS written information is different (either consistent or conflicting) from that in the census. How to use the table is illustrated by the following verbalization of the symbols of "Situation 1" (see the first column of the table). Situation 1 states that if the person's written PMS entry is essentially the same as the census entry, and if the written entry in question 22 indicates that the person belongs in a PMS category other than the one associated with the code entered in the code box of question 22, and if the code entered is 459 or 437-448, then, in spite of the written entries in PMS questions 23 and 24, the PMS code is to be treated as an essential part of the PMS occupational description. Conditions and Actions Situations Conditions The written entries in the PMS convey essentially the same information as the census written entries , The written entry in question 22 conflicts with the numeric code in question 22 The written entry in question 23 conflicts with the numeric code in question 22 , The written entry in question 24 conflicts with the numeric code in question 22 The numeric code in question 22 is 459 or 437-448 The numeric code in question 22 is 451 Actions Treat numeric code in question 22 as additional information... Ignore the numeric code in question 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 NA Y Y Y Y Y Y N N N N N N N N Y N N N Y Y N N N N Y N * Y N N * * Y Y N N * Y * * Y N * * * •k Y Y Y Y N N N NA NA NA NA NA NA NA NA NA NA NA NA NA Y N Y N Y N X X - - - X - X - X - X _ _ X X X _ X _ X _ X _ NA 1 For conditions: Y - means "yes", the condition must be true. N - means "no", the condition must be false. * means ignore, the condition is not relevant. NA - means "not applicable". 2 For actions: X means execute, i.e., perform the action. means ignore, i.e., do not perform the action T2 APPENDIX C Census Rules for Coding Occupation Note: Excerpted from U.S. Bureau of the Census, 1970 Decennial Census of Population, Procedures Manual, Volume III, Part V, Chapter A, "Industry and Occupation Coding", May 6, 1970. After an industry code has been entered in item 33, examine items 34a, b and c. From this find the appropriate code in the Alphabetical Index following the instructions below: The final determination of a correct code for a particular listing in the occupation portion of the Index is the result of the proper consideration of the occupation return (item 34a, b, c), the code of the industry return (item 33a, b, c) and sometimes the class of worker item (item 35). If a written entry cannot be coded after following the Index instructions, refer the entry. a. Coding to most specific entry. In determining the proper code for an occupation return, consider the entries in items "a" and "b" as a combined entry. For example, if item "a" says "machine operator" and item "b" says "runs a lathe," combine the entries and code "lathe operator" not "machine operator" because it is the most specific entry. On occasion, two distinctive jobs will be described in "a" and "b"; for example, on line "a" "receptionist-typist" may appear and on line "b" "typist" will be written. In such a case code that job appearing on line "b", for the respondent has told us he considers that his main activity. Sometimes line "b" repeats the double job function given on line "a", for example "receptionist-typist" appears on line "a" and also on line "b". In such a case code the lowest number code, 1 in this case code 364 for receptionist. If one of the codes in the comparison is a letter code, use the following numeric equivalents of the letter codes. In coding, use only the letter code, never their number equivalents. N-142 S-473 W-801 P-305 T-602 X-903 Q-372 U-715 Y-915 R-415 V-751 Z-984 At times the Index lists activities along with a job title which means you must refer to Item 34b of the schedule. For example, if 34a says "porter" you will need to look at 34b for clarification. If 34b says "cleaning" or the like, it is covered in the Index listing of: Porter, cleaning 902. The Index also has a listing for: Porter, baggage 934, which would be used if item 34b had said "handle baggage" or a similar entry. b. Additional rules for use of items 34b and 34c (1) When not to use item 34b. — At times the Index will say "any activity." In this case you will not use item b. If, for example, 34a says "secretary" and 34b says "filing and typing," you will find in looking up secretary, the Index gives a listing for "Secretary, any activity." For this entry, you will use only 34a and disregard the activities listed in 34 b. (2) Use of job title question (34c) — Item 34c should be used only when a code cannot be assigned by using 34a or b, for example, where both 34a and 34b are too broad or general. In such a case, code using 34c if it clarifies and/or is consistent with items 34a and 34b. If it is not consistent, use the rule of lowest number in instruction 3a. c. Coding of occupation from Industry. In some cases the information found in all parts of item 34 is not enough to properly code occupation, but the needed informa- tion can be found in item 33. For example, item 33b may say "auto body repair shop" and item 34 "mechanic." In this case you can code the person as an auto body mechanic. d. Alphabetic Index. Occupation titles are listed in the Index in several ways. These are: (1 ) Occupation titles with no restrictions Ticket writer 394 This means that if the schedule entry reads "Ticket writer" the proper code to be entered in code box 34 would be 394. (2) Occupation titles with Industry and/or Class of Worker restrictions . There are a number of types of restrictions. In all cases the occupation title is in the left columns and the occupation code in the right column. (a) Title with one industry code in the center: Salesman 287 705 This means that you can code your occu- pation 705 only when the industry has been assigned the code 287. (b) Title in combination with range of industry codes 1 This rule is known as the "lowest-code" rule. Compensator man . . . .307-318 620 If the industry code had been assigned 308, 34 35 you could code the "compensator man" 620, because code 308 falls within the range 307-318. If the industry code does not fall within this range of codes in the middle column, you cannot use the code for occu- pation. (c) Title in combination with several industry codes: Criminal investigator. . .L, M, 907, 927. . .964 In identical fashion as the industry range, here the occupation code 964 can be used only if you have assigned as the industry every entry on your schedule one of the four codes in the center. APPENDIX D Influence of Professional Indentification on Reporting of Postcensal Manpower Survey Occupation The table in this appendix was created in the hopes of indicating the influence of professional identification in 1972 on the PMS response when it conflicts with the census response. The universe for the table is restricted, therefore, to cases with conflicting response differences. ANALYSIS In the majority (about 61 percent) of the conflicting responses, the respondent reported a professional identification 1 in 1972 in the PMS that was the same as the 1970 PMS occupation. In 1 Item 41 of the PMS questionnaire asked the respondent to complete the following statement by inserting a code and a description from List C: "Based on my total education and experience, I regard myself professionally as a (an) . . ." only about 16 percent of the cases did the respondent report a professional identification that agreed with the 1970 census occupation. In nearly 18 percent of the conflicting-response cases, the person gave a professional identification that did not match his or her 1970 census occupation or his or her 1970 PMS occupation. The 1972 professional identification was not reported in about 6 percent of the cases. The percentage of conflicting-response cases for which the professional identification agreed with the 1970 PMS occu- pation differs very little among the various occupational groups. The percentage ranges from about 59 percent for life scientists to about 63 percent for mathematical specialists. There is more variation (about 8 percent for social scientists to about 27 percent for life scientists) among the occupational groups for persons whose professional identification matches their 1970 census occupation. Table D. Conflicting Responses, by Agreement Between 1972 PMS Professional identification and Detailed 1970 PMS and Detailed 1970 Census Occupation, by Major 1970 Census Occupational Groups Total Professional identification in 1972 Same as 1970 PMS occupation Different from 1970 PMS occupation Major 1970 census occupational group Total Same as 1970 census occupation Different from 1970 census occupation 1972 Professional identification not reported ALL OCCUPATIONS 290 100.0 67 100.0 67 100.0 38 100.0 34 100.0 61 100.0 25 100.0 176 60.7 40 59.7 41 61.2 24 63.2 20 58.8 38 62.3 15 60.0 97 33.4 22 32.8 23 34.3 11 28.9 12 35.3 21 34.4 9 36.0 46 15.9 8 11.9 12 17.9 4 10.5 9 26.5 12 19.7 2 8.0 51 17.6 14 20.9 11 16.4 7 18.4 3 8.8 9 14.8 7 28.0 17 OPERATIONS AND COMPUTER SPECIALISTS 5.9 5 ENGINEERS 7.5 3 4. 5 MATHEMATICAL SPECIALISTS 3 LIFE SCIENTISTS 7.9 2 5.9 PHYSICAL SCIENTISTS 2 3.3 SOCIAL SCIENTISTS 1 4.0 Note: Agreement between 1972 PMS profes detailed level of occupational classificati sional ide on. ntif ication and PMS an d census occu pations was dete rmined at the 37 APPENDIX E Reconciliation Process The classification of the causes of mismatches is described in the body of the report. The process by which each mismatch was placed in one of these categories is described in this appendix. Throughout this process, a hierarchy of causes was established; the lower the category appears in the classification scheme shown in example 2 (see page 8), the higher it is in the hierarchy (for example, "methodological differences" are higher than "PMS coding errors"). A search was made for that cause in whose absence the PMS and census occupational categories would have matched or corresponded. In most cases, if two or more reasons for the mismatch were identified, the highest one in the hierarchy was considered to have caused the mismatch. The first step in the process was the independent examination of both the census response and the PMS response. The basis of the code assigned to the census response was examined, and, for cases in which the assignment depended upon an industry or class-of-worker designation, the bases of these codes were also examined. If a census coding error was discovered, this error was considered to be the sole cause of the mismatch, and no attempt was made to locate another cause. This procedure was the only exception to the rule that the highest of two or more reasons was chosen as the only reason for the difference. If the census coding was correct, the PMS response was examined for coding and processing errors. After an independent verification of each report was made, the PMS response was compared with the census response. Cases with the same or essentially the same responses were separated from those with different responses. The former cases were then closely examined for structural or methodological differences. Finally, cases with different responses were investigated. The consistent responses were separated from the inconsistent or conflicting ones. The consistent responses were examined for evidence of the various kinds of insufficient responses; for the conflicting responses, an attempt was made to find the reasons for the conflicts. Discovering the underlying reasons for the conflicting responses put the investigator into an area of speculation, and such factors as the reference periods, the company names and locations, and the person's professional identification (see appendix D) for certain jobs were examined. The chart in this appendix is a flowchart of the entire reconciliation process. 39 Enter a mismatch Examine the census response CAUSE OF MISMATCH Yes Census coding error Examine the PMS response Compare PMS response with census response No Yes Compare PMS response with census response PMS coding or processing error Yes Structural difference: PMS subcategory misclassification Compare PMS response with census response No Structural difference: PMS residual - category classification Yes Methodological difference: managerial concept No Methodological difference: other Yes Insufficient response 40 Conflicting response APPENDIX F Sample Design, Estimation Procedure, and Reliability of the Estimates SAMPLE DESIGN AND ESTIMATION PROCEDURE The sample of occupational classification differences (mis- matches) was a stratified, systematic sample of mismatches selected from a universe of mismatches consisting of all scientific and technical occupational classification differences between the 1970 census occupational classification and the PMS occupational classification. The universe was stratified by major occupational groups and by level of mismatch, forming a total of 12 stratums. The sample size in each stratum was determined so as to produce a coefficient of variation of at most 12.5 percent. The estimates produced from the sample of mismatches are attributes in the form of proportions. The standard errors are estimated assuming the systematic sampling procedure is equi- valent to a simple random sample of mismatches. RELIABILITY OF THE ESTIMATES The sample used for this match study is only one of a large number of possible samples of the same size that could have been selected using the same sample design, sample selection. and measurement procedures. Estimates derived from these samples would differ from each other. The standard error is a measure of the variation among the estimates from all possible samples and is, therefore, a measure of the precision with which an estimate from a particular sample approximates the average result of all possible samples. The estimate and its associated standard error may be used to construct a confidence interval; that is, if all possible samples were selected, each of these surveyed under essentially the same general conditions, and an estimate and its estimated standard error were calculated from each sample, then approximately 95 percent of the intervals from two standard errors below the estimate to two standard errors above the estimate would include the average value of all possible samples. The average value of all possible samples may or may not be contained in any particular computed interval. But for a a particular sample, one can say with specified confidence that the average of all possible samples is included in the constructed interval. Simi- larly, the chances are about two out of three that the survey estimate will differ from the average result of all possible samples by less than one standard error, and 99 out of 100 that the survey estimate will differ from the average result by less than 2Y2 times the standard error. 42 **************************** Now available and filled with current and historic economic, political, and social statistics POCKET DATA BOOK U5A1976 represents the fifth edition in this continuing series of popular, compact, and portable reference books. In commemoration of the Nation's Bicentennial, a special section, "Bicentennial Statistics," is included with historical tables and charts which summarize the growth and changes in the U.S. since 1776. Illustrated historical charts along with current information are also featured throughout the book on issues of daily interest. 444 pp. $4.00 Copies may be purchased from Superintendent of Documents, Government Printing Office, Washington, D.C. 20402 or from Department of Commerce District Offices or GPO Bookstores located in major cities. A 25-percent discount is allowed on orders of 1 00 or more copies sent to the same address. You may also use the convenient order form below. of /- & ay J CJ ■ o a POCKET DAT *~'m A BOQK U SA19-7B (please detach along this line) ORDER FORM Please send me copy(ies) of POCKET DATA BOOK, USA 1976 $4.00 (S/N 003-024-01155-7) City I enclose $ coupons) or charge to my Deposit Account No Total Amount $ MAIL ORDER FORM WITH PAYMENT TO Superintendent of Documents or any U.S. Department of Commerce District Office (check, money order, or Supt of Documents MAKE CHECK OR MONEY ORDER PAYABLE TO SUPERINTENDENT OF DOCUMENTS To Insure Prompt, Accurate Shipment, Place Correct Address on Mailing Label Below Superintendent of Documents U.S. Government Printing Office Washington, D.C. 20402 Penalty for Private Use, $300 Street Address POCKET DATA BOOK, USA 1976 City, State, and ZIP Code For Office Use Only Quantity Charges M, I To Mail Sub Coupons Handling . . - POSTAGE AND FEES PAID U.S. GOVERNMENT PRINTING OFFICE 375 Special Fourth Class Rate Book 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- 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. SUBSCRIBER SERVICES SECTION (ASD) BUREAU OF THE CENSUS WASHINGTON, D.C. 20233 Superintendent of Documents U.S. Government Printing Office Washington, D.C. 20402 Official Business Postage and Fees Paid U.S. Department of Commerce Special Fourth-Class Rate- Book COM-202 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 Guide for Local Area Population Projections Current Population Survey: Design and Methodology An Evaluation of 1970 Census Occupational Classification: The Postcensal Manpower Survey- Census Match Study To receive further information or to order copies of these reports, contact the nearest district office of the Department of Commerce or write to- „.,„., 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 No. 41 SUBSCRIBER SERVICES SECTION (ASD) BUREAU OF THE CENSUS WASHINGTON, D.C. 20233 A0000147b| D S7E