College and Research Libraries Online Catalogs and Their Users Robert N. Broadus A review article on the CLR study of online catalogs. he current interest in research on computer catalogs is fortu- nate at this stage of their devel- opment when so few libraries have them. Not only can other librarians, if they are smart, take advantage of the ex- periences of pioneers, but a good deal of standardization can be postponed until more knowledge is gained. Who knows what card catalogs would be like today had they been studied with such diligence in the latter part of the nineteenth cen- tury? During April and May 1982, question- naires were administered to users and nonusers of online catalogs in twenty- nine libraries across the United States. The emphasis was on ''the interaction be- tween the human user of the computer catalog terminal and the computer system which supports and responds to the us- er's request for information" (p.30). * The research was sponsored by the Council on Library Resources; but five other organi- zations were involved, each preparing its own report pertaining to a group of li- braries and/ or offering general interpreta- tions. Four of the reports, 1 each covering a dif- ferent set of data and representing a differ- ent emphasis, are summarized and syn- thesized in Matthews' report, Using Online Catalogs, which is the focus of th.is review. The fifth organization, Online Computer Library Center, Inc. (OCLC), administered questionnaires in ten li- braries, five of which used OCLC termi- nals as catalogs. OCLC submitted a three- volume ''Final Report to the Council on Library Resources," which was received after this article was substantially com- plete. The Matthews summary includes data from the questionnaire survey in nine of these libraries (one was dropped) but not interpretations given in OCLC' s re- ports.2 SCOPE AND METHOD The specific goals of the survey were: 1. to produce data for analysis that would enable designers of public com- puter catalogs to improve computer cata- log system interface features. The system interface includes commands, displays, indexes, and similar software- and hardware-related features. 2. to gather data and prepare an analy- sis that would enable libraries to improve the implementation and support services for online public access catalogs. 3. to collect additional data that would enable libraries to extend public access computer catalog services to patrons who were not yet users (p.8). The twenty-nine participating institu- tions included two state/federal libraries (one being the Library of Congress), four- teen ARL libraries (including nine cam- puses of the University of California with its online union catalog), seven other aca- demic libraries, two community colleges, and four public libraries. Not only were *Page numbers in parentheses refer to the summary report: Joseph R. Matthews and others, eds., Using Online Catalogs: A Nationwide Suroey, A Report of a Study Sponsored by the Council on Library Re- sources (n.p.: 1983). Robert N. Broadus is professor, School of Library Science, University of North Carolina at Chapel Hill . 458 there many types and sizes of libraries, sixteen different computer systems were investigated (p.16). The principal means for gathering infor- mation about these phenomena were printed, self-administered question- naires. One questionnaire had fifty-nine questions for persons who had just used an online catalog; another had fifteen for those who, though encountered in the H.: brary, said that they had never used such a catalog. All library staff members were excluded. The University of California ad- ministered the questionnaire online, "but the text of the questionnaire and the method of administration parallelled those used at other locations. " 3 This li- brary system did not question nonusers. The questionnaires evidently were pre- pared and administered with great care. Planning began in December 1980, and a pilot survey of three thousand subjects was conducted in 1981. In the forty-nine- page Data Collection Manual, precise in- structions were given to study coordina- tors, supervisors, and data collectors in each library. Included were scripts for use in explaining various matters to the sub- jects. There is an advantage in bringing so much expertise to bear from different sources on one study. The fact that all us- ers responded to the same set of questions (and all nonusers to another) gives the study a unity and power not achievable by dozens of piecemeal approaches. The dis- advantage, however, is that the question- naire did not fit all libraries equally well. One may wonder whether a respondent was affected by question 3, ''I searched for what I wanted by . . . A subject heading or headings," if that particular catalog of- fered no such option. It may have been distracting to see question 26, ''Using logi- cal terms 'like AND, OR, NOT is difficult ... Strongly agree ... Agree," in those fifteen libraries that did not offer Boolean searching. Another problem is whether respon- dents understood all the questions, even those applicable in their own libraries. Ap- parently a fair number of subjects dealing with question 46, "Select up to FOUR kinds ,of material you would like to see Online Catalogs 459 added to the computer catalog," mistook "Journal or magazine titles" for titles of articles. 4 At Northwestern University,· 15 percent of the users checked "Journal ti- tles" as a kind of material they wished to see added to the database, though the li- brary already had practically all its period- ical titles in the computer catalog. 5 Did all people responding to question 3, "I searched for what I wanted by . . . ,'' dis- tinguish "A topic word or words" from "A subject heading or headings"? Of those users in systems with no online aids, 28.5 percent said they had received help from instructions on the terminal screen, and the wry comment from the University of California is justified: ''One might wonder just what 'instructions on the terminal screen' were used by [these persons]. ''6 On the other hand, the RLG report, questioning the number of people ~ho claimed they had received no such help, points out that practically all users necessarily got some "minimal assis- tance" from the screen. 7 Undoubtedly there was a problem with the question- naire's meaning here. The Data Collection Manual instructed surveyors to respond, "Please just decide what you think it means, and _answer accordingly," when told by subjects, "I don't understand what this question means.'' 8 · The original goal of the combined sur- veys was to collect information from 25,000 persons, 9 but only 13,591 users and 7,625 nonusers were approached (p.199a) and 8,094 users and 3,981 nonusers responded-still a respectable number. Well over half the questionnaires were col- lected from ARL libraries, about one- fourth from public libraries, 8 -percent from other academic libraries, 7.5 percent from state or federal libraries (almost all of these representing the Library of Con- gress), and only 3 percent from commu- nity college libraries. The numbers of questionnaires completed at various insti- tutions differed widely. The Ohio State Li- brary furnished only five for users and eighty-six for nonusers. The University of California, because of its online survey, contributed no nonusers. Four of the twenty-nine libraries accounted for fewer than fifty users each. 460 College & Research Libraries The overall rate of response (percentage of those who, when approached, agreed to complete a questionnaire) was 59.6 per- cent for users, and 52.2 percent for nonus- ers (p.199)-not a bad rate as surveys go, but the investigators are aware that some results could have been distorted by the failure to obtain answers from many other people. This response rate differed from library group to library group. For the six analyzed by Matthews, the rate was 46.5 percent for users; 62.2 percent for nonus- ers.10 Furthermore, there was remarkable variance, apparently, from library to li- brary. At Syracuse University, of the users approached, a whopping 93.9 percent agreed to work on a questionnaire; of non- users a mere 24.1 percent (p.199, 199a). However, there is a discrepancy between these figures and those in the OCLC re- port.11 Not all questions were answered by every subject. At the University of Califor- nia many of the items were passed over-a cause for concern as expressed by David Bishop. 12 In fact, of the 8,094 users' ques- tionnaires, only 2,501 included all the thirty-one answers to part II, which dealt with "your experience with computer cat- alog features" (p.202). The study's validity was measured, to some extent, by checking questionnaire answers against records of users' behav- ior. The records were indicated by the computer transaction logs provided by some of the participating libraries. For in- stance, in the California system, answers to the question about type of search made (author, subject, etc.) agreed reasonably well with transaction records. For exam- ple, 15.2 percent of the users said they had used an author search; 44.5 percent said they had used a subject search. The log in- dicated that these percentages were 12.0 and 43.8, respectively .13 There are prob- lems, however, in a comparison of the transaction log at Northwestern and their questionnaire responses. According to the log, 38.4 percent of the access points used were Library of Congress subject head- ings and 25.1 percent were author head- ings.14 However, the search information reported on the questionnaire was: com-· November 1983 plete author, 23 percent; part author, 11 percent; subject heading, 17 percent; and · topic word(s), 10 percent. 15 Here the logs represented not the same individuals who responded to the questionnaire, but those who used the catalog a month or two be- fore the survey. It should be noted that transaction logs can record all users ( dur- ing the time period covered); no one can refuse to be "questioned." Unfortu- nately, it was not possible to check re- sponses to many questions in this way. However, the results obtained, while rais- ing doubts, do bring some measure of as- surance, as the same general findings seem to occur with regularity in all five of the group reports. · FINDINGS AND APPLICATIONS To focus more on what the survey dis- covered and how the material was ana- lyzed and interpreted, the fi,rst thing to note is the tremendous problem faced by the investigators. With so many libraries, catalog features, and questions, some of them having various response options, there are millions of bits of information that could have been brought out and ex- amined. The analyses and presentations are, on the whole, admirable. This article will examine some of the findings in roughly the same order as they are given in the Matthews summary, and then will indicate some unsatisfied curiosity. First, though, the important overall con- clusion: "To say that users have very posi- tive attitudes towards the cqmputer cata- log is perhaps the understatement of the year'' (p .139). ''Over ninety percent of us- ers like [it]" (p.140). "The evidence shows that those who have used the computer catalog love it, and those who have not used the computer catalog like it almost as . well" (p.176). While these statements are justified per- haps by the fact that 67 percent of the us- · ers said their attitude toward the catalog was very favorable and 25 percent some- what favorable (p.141), we might feel more comfortable if the exuberance were toned down. What of the 5,500 persons who, having just used the online catalog, refused to cooperate? While it is easy to imagine a user, angry at failure, wanting a chance to express that on paper, it would seem more likely that those willing to take twenty minutes or so to fill in the blanks were in a better mood than were those who refused. For some· users, the act of re- jection could have been a sign of negative attitude toward the whole situation. Also, the acquiescence bias-the ten- dency to say yes as a way of not seeming disagreeable-may well have affected re- sponses to some of the questions. Gener- ally in surveys this factor is hard to evalu- ate. Though it is supposedly less serious on a written questionnaire than in face-to- face interviews, it cannot be discounted entirely here. Two other points bear on this matter. Data collectors were told in the Data Collec- tion Manual that they should not approach users when the catalogs were out of ser- vice, because these people "will have a bad impression [of the catalogs].'' 16 Also, the RLG report, in addition to analyzing questionnaire results, gives quotations derived from interviews with library staff members at Stanford, Northwestern, and Dartmouth, not mentioned in Matthews' summary. The tone here is less positive. 17 The overall conclusion on favorability, then, would seem a bit exaggerated, though it is virtually indisputable that most people like the online catalog. It is unfortunate that the report makes so few direct comparisons between those who have used and those who have not used the new catalogs; some of the differ- ences between the two groups were not so great as may be implied by a few of the conclusions. Thus, the statement that us- ers of the online catalog are also frequent library users (p.44) should be viewed in light of the fact that even of the nonusers, a full 59 percent claimed that they visited the library daily or weekly (p.55). The interpretation of findings regarding male and female subjects is not complete. Under the subheading "Computer Cata- log User Is a Young Adult" (original in italics, as are all subheadings) is the sen- tence, "The typical user of the computer catalog is male (60% of the users were Online Catalogs 461 male)" (p.48), improved somewhat by the observation (p.56) that ''women are some- what less likely to be users of computer catalogs than men. While women make up only 40% of the user population, they comprise 48% of the non-user group." The relatively low amount of online cata- log use by females (only 22.5 percent at UCLA and only 33.1 percent in the entire University of €alifornia system)18 is some- what confusing. However, it would seem relevant also to state the findings in this way: Of all the females who responded to the question about their sex, 61.2 percent were users; 38.8 percent, nonusers. Of the males, 68.5 percent were users; 31.5 per- cent nonusers. 19 This information is bal- anced by the comment in the OCLC report that in the Syracuse, Ohio State Univer- sity, Dallas Public, and OCLC system li- braries, females outnumbered males as firsttime users, a sign that females are be- ginning to use these catalogs more. 20 The subheading "Computer Catalog Users Are Highly Educated" is backed up by the fact that 90 percent said they had completed some college work or were col- lege graduates (p.48), but surely that was to be expected in view of the great number of respondents in college and university li- braries. Table 7 (p.57) shows that for non- users the corresponding percentage was even higher, at 91-a fact ignored in the text. The University of California report, Users Look at Online Catalogs, presents in- formation about users and nonusers in parallel columns. These tables must be read carefully or they may be misleading, but they do give the facts in more conven- ient form. ''There is no evidence of an a priori bias against use of the computer catalog by any disciplinary group. Users from the Arts and Humanities, Social Sciences, and Business Management represent 57% of all respondents" (p.49). Under the sub- heading "Academic Non-users Similar to Academic Users" (p.58) is added, "More non-users of the computer catalog come from the following disciplines: Business/ Management, Arts and Humanities an<:f. · Engineering.'' A better· picture is pre- sented in table 16 of the University of Cali- 462 College & Research Libraries fornia report, 21 which reads in part: Arts and Humanities Physical/Biological Sciences Users Nonusers 77.5% 22.5% 66.6 33.4 Social Sciences 74.0 26.0 Business/Management 55.0 45.0 Engineering 61.3 38.7 The table is to be interprefed: of all the Arts and Humanities people who an- swered questionnaires, 77.5 percent were users of the online catalog; whereas of all the Business/Management people, only 55 percent were users. This is a somewhat different impression from that given by the text of the summary volume. It still might be a mistake to interpret these per- centages as if the persons answering (us- ers and nonusers) constituted an adequate sample of the population who visit these twenty-nine libraries. Those who frequently used other com- puter systems were not much more likely to consult the online catalog (p.46,47) or to express satisfaction with it than the gen- eral run of library clients (p.172). In fact, those who never used other computers had a somewhat lower error rate, accord- ing to the University of California transac- tion log, than those who used them fre- quently.22 Here is a disconcerting observation about the reporting of the study: the user questionnaire for the pilot study had ques- tion 72, ''I use this library's book, card, or microfilm catalog: a. Every visit b. Almost every visit c. Occasionally d. Rarely e. Not before today.'' Response e. was a little odd in that it implied that the number who used the conventional catalog ''today'' for the first time might be substantial. The line was changed in the final (main) user questionnaire, where question 50, re- sponse e. reads, "Never." Qu·estion 8 of the nonuser questionnaire keeps "Not be- fore today'' as a possible choice on the same item, and apparently the University of California retained it for its online user survey. In table 3 (p.46) of the Matthews summary report, however, there are re- corded percentages for users who, on question 50, are said to have indicated, "Not before today." The same occurs in November 1983 several of the intermediate reports (though RLG, in table 25, has the response as "Never"). It may be that the error is merely a harmless mislabeling, but be- cause of the wording of the pilot question- naire and question 49 of the final user ver- sion (which includes "Not before today"), the reader is left with an uncom- fortable feeling. The subheading "Non-users Say They Like to Use the Computer Catalog" (p.61) when read ". . . are Likely to Use . . . " is better understood in light of the fact that nonusers of the online catalog don't con- sult other forms of library catalogs very much either, a point strengthened by the anal~sis in a University of California re- port to which passage the summary vol- ume refers. Organizational Setting and Computer System Answers relating to the systems and set- tings indicate that users are generally sat- isfied with the way libraries have pro- vided facilities and instructions for utilizing the new catalogs, though 51 per- cent feel that their respective libraries need more terminals (p:77), and 34 per- cent want more writing space (p.74) or, perhaps, more printers. A confusing fact is that users seem to rely more on library staff for help if one on- line aid feature (as opposed to none) is available; whereas two online features of this kind (say both instructions and HELP) seem to reduce the need for such aid (p.72). Response time of the system was a prob- lem, especially for users of public libraries, for 30 percent of all users (p.88,89). Most of them preferred the waiting times to be more uniform, rather than a delay of about eight seconds followed by one of forty. Thirty percent wanted newspapers added to the catalog coverage (p. 93), but possibly some of them were thinking about index- ing rather than mere titles. Government publications, older books, dissertations, and "journal titles" (articles?) were also requested, each by more than 20 percent of the respondents. Answers to this ques- tion are difficult to evaluate because the content, clientele, and size of database dif- fered so much from library to library. The ''Human-Computer Interface'' was examined by questions relating to twenty- seven features of the catalog, calling for Likert-type answers ("strongly agree, agree, neither, disagree, strongly dis- agree, does not apply"). Respondents ei- ther agreeing or strongly disagreeing are grouped together in table 17 (p.101). Al- though the positive/negative orientations were alternated for the actual questions in this section of the instrument, table 17 points up the problems by showing, for each question in this group, the percent- age of responses in the two (of five) cate- gories least favorable to the online catalog. Users' perceptions of problems did not differ greatly by type of library (p.101). The problem most often (46 percent) checked was, "Increasing the result when too little is retrieved ... '' Forty-three per- cent agreed that ''Finding the correct sub- ject term is difficult,'' and 37 percent had trouble remembering what ·was included in the computer catalog. Thirty-one per- cent of the users agreed with the state- ment, ''A computer search by subject is difficult," but most of them expressed no difficulty with searches by author, title, or by a combination of the two. The summary provides more in-depth analysis by showing what problems seemed to occur in relation to the features offered by the sixteen different systems. For instance, the difficulty of increasing the result seemed less severe where online printing was available; more of a problem in those systems that were menu-driven (p .106). As the authors point out, the great number of variables make such analysis extremely difficult, and the results possi- bly misleading (p .103) because the "evi- dence is not strong enough to demon- strate clear superiority of any particular feature" (p.111). The full values of the sur- vey, however, can be realized only by such attempts. Also useful is the discussion of design trade-offs (p.11~), e.g., a feature may make a certain kind of search easier or more effective, yet so add to the complex- ity of the catalog that more users find it confusing. It would be extremely helpful if librari- Online Catalogs 463 ans could learn what added features would be most useful to users. Such infor- mation is unusually dtfficult to obtain be- cause average laity have no way of know- ing what they need/want until they try it, or until it is at least suggested to them. The questionnaire method is not very effective for gathering ideas of this kind, but the in- vestigators made a good attempt by pro- posing fourteen possible features, plus the category "None." Even here, the question must be raised: Were respon- dents aware of what it would be like to "search by call numbers" or to "search a book's table of contents, summary or in- dex"? The latter was ranked second high- est, in any event, after "Ability to view a list of words related to my search words.'' About one-fourth of the respondents checked "Ability to know if a book is checked out,'' but only 10 percent wanted to know the location of books in the library (p.114). (The University of California and Library of Con~ess reports give 15.1 as this percentage.t Only 18 and 24 percent, respectively, desired ability to search by ti- tle word or by subject word. Analysis of these data, leaving out responses ob- tained in those libraries already offering the proposed feature or features, is of- fered by a University of California docu- ment, and the percentages for most of the features noted above are thus a little higher. 25 The Matthews summary volume also uses factor analysis to bring together in groups those features that seem to cause problems for users. The seven factors de- rived are helpful in giving a clearer picture than the raw data. Catalog Use and Satisfaction In those catalogs that provided subject access, about 59 percent of users searched for subject information (p.129), a finding not surprising in view of the great number of undergraduate students responding. In community colleges, this percentage was higher. Searching by keyword was em- ployed heavily when available (though not missed much where it was not); key- words in subtitles were frequently utilized (p.136). About 85 percent of all users found 464 College & Research Libraries some or all of what they had been looking for (p.138), including 17 percent who found more than they had intended. Though 16 percent didn't find what they were seeking, about half answered yes to the question, "I came across things of in- terest other than what I was looking for." Serendipity did not seem to be related to the proportion of a library's hoiUings that had been put into the database. Another important finding was satisfac- tion rate. Responding to the statement ''In relation to what I was looking for, this computer search was . . . ," 46 percent marked ''Very satisfactory'' and 34 per- cent, "Somewhat satisfactory." Overall, these users, especially the younger ones, expressed a general liking . for the com- puter catalog and seemed quite positive that it was better than the card, book, or microfilm catalog. Even those who had just finished an unfruitful search had high opinions of the online catalog (p.140-42). In the group of libraries surveyed by OCLC, 50 percent of the nonusers who said they did not like to use computers ex- pected to use the online catalog in the future 26 -surprising in view of the fact that 43.6 _percent of the nonusers in the aggre- gate said they consulted the traditional catalog rarely or "Not before today" (p.55). There were discrepancies that may prove meaningful when checked against the differences among systems. On the University of California campuses, 70.8 percent used ''Very favorable'' to describe their attitudes toward the computer cata- log, against 67.0 percent for the total re- spondents in the study. However, com- paring the computer catalog with a manual one, only 68.3 percent of the Cali- fornia users considered the former better, whereas overall, 74.5 percent did. ("Can't decide" was excluded from this tabula- tion. )27 What variables, as indicated by. respon- dents, were related to success or failure in searching? While the investigators are not able to answer decisively, they often make good attempts. Some of the findings were far from sensational; for example, diffi- culty in subject searching corresponded to a lesser amount beingretrieved (p.146). On the other hand, it was startling that us- November 1983 ers found less of what they were looking for in those systems that allowed subject searches (p.148). The libraries not using OCLC as a catalog were the ones that al- lowed such subject searches, and the available databases in these libraries were generally smaller than in those libraries using OCLC. The researchers speculate, therefore, that it was really the smaller size of the databases that caused the fail- ure to retrieve greater amounts of material by subject searches. Yet, Ohio State Uni- versity, which offered subject searching, had a large database (about 1,500,000 ti- tles), and some of the other "subject search libraries" had sizable bases. It should be noted that only 329 ( 4.1 percent) of the total user questionnaires were com- pleted at libraries that did not offer subject searches, 28 so they could not have affected seriously the overall outcome. Taking all twenty-nine libraries into account, subject information searches were 56.5 percent of all searches. When the five libraries not providing subject searching were omitted from consideration, the percentage of sub- ject searches in the remaining libraries was not much higher, 58.8. 29 Surely this important matter calls for more analysis, as does another curious finding: difficulty in understanding displays was associated, though not strongly, with retrieval of more wanted information (p.149). What variables were associated with the users' perceptions of searches as satisfac- tory or unsatisfactory? Again, some results are obvious. For instance, where there was a higher degree of satisfaction, fewer problems had been encountered with subject searches (p.151), but both sat- isfaction with search and favorable atti- tude toward the catalog generally were re- lated negatively to a system's allowing searches by local accession numbers (p.153, 160). Also, ability to search by call number was related to a lower satisfaction level (p.162). Perhaps the problem was that users were further confused by one more option, as the authors suggest. This is another point that needs further explo- ration, especially in view of the fact that the provision of unusual access points was associated with greater favorability to- ward the catalog (p.161). One report makes the comment that "it is surprising that more users did not encounter more problems with a greater number of the available computer catalog features.' ' 30 In libraries providing for combined au- thor/title searches, users tended to report finding more information than they had been looking for (p.l55). The reason, as the authors note, is difficult to see, but even more difficult to account for is that users who had problems understanding displays on the screen also were more likely to find more than they were looking for (p.157). There was some tendency for users of those systems without subject search ca- pabilities (the OCLC catalog group) to ex- press even greater general satisfaction with the online catalog (p.161), and to fa- vor it more strongly over the card catalog (p.165). Again, the only explanation of- fered is the larger databases available to. these libraries, but the explanation is hardly sufficient here either. The finding could be related to the difficulty of subject searches in general. The OCLC analysis of tapes from three libraries not using OCLC as a catalog found that ''subject searches result in zero hits more often than other searches, with the exception of a DUAL search (Boolean) at Syracuse. " 31 In the libraries that included formats other than books and serials in their online catalogs, users were less likely to express a preference for the catalogs (p.l66). The ev- idence is not strong, and certainly does not mean that libraries should now decide to exclude non print materials from cover- age in online catalogs. However, it does suggest further investigation. Also hard to explain is that the capability of limiting search results by language of publication corresponded to a slightly negative attitude toward the online cata- log (p.l66), as did the capability of search by series title (p.167). Not appearing to have been related to satisfaction were such features as key- word or Boolean searching (p.172). But, here again, the results are clouded be- cause so many of the systems tested did not offer these capabilities. It is also possi- ble that because features like these are not found in traditional catalogs, users were Online Catalogs 465 not accustomed to them and did not even try them. In the future, with more users taking advantage of such features, there will be better opportunities to test their utility. Implications The last chapter of the Matthews sum- mary report draws practical conclusions, directed at library managers, reference de- partments, systems designers, and the li- brary profession as a whole (p .175). In ad- dition to urging provision of more online catalogs, this chapter stresses such needs as placing catalogs where users "can't miss them'' (p.176) and improving subject access. Important with respect to the latter is the capability of allowing users to search by keyword of subject headings and titles, and to browse the subject list or a thesau- rus. There is a good deal of comment in the various reports about the possibility of augmenting the bibliographic entry with the work's table of contents and/or index, and of making these terms access points for searching. Two serious factors for con- sideration, however, are costs and the possibility of confusing clients even more. Briefer bibliographic records are, accord- ing to some indications, about as satisfac- tory as full entries-a finding in ar:eement with the recent article by Seal. 3 One re- port suggests the provision of optional displW,s for those people desiring full en- tries. If additional elements are added to the bibliographic record, deciding which of them should be access points will be a problem. 34 Making it easier for users to give com- mands would also seem to help, as would providing them with the opportunity to determine whether a given book is avail- able in the library (p.183). A FINAL WORD Among the many unanswered ques- tions are: How do responses of those us- ing touch-screen systems, as at Evanston and Iowa City public libraries, compare with those from public and other libraries with keyboard systems? Another: 8 per- cent of the aggregate sample of users indi- cated that they found the terminal table ei- ther too high or too low. The University of 466 College & Research Libraries California traced the problem primarily to the Irvine and Berkeley campuses, 35 but they did not reveal the heights of the satis- factory and the offending tables. In the printed questionnaire survey at various li- braries, a few facts were noted about each person who, when ap}iroached, refused to answer questions -approximately . nine thousand in all. So far the results have not been published. The final summary does not include analysis by individual library, but fortu- nately the intermediate reports do. Those by the University of California, RLG, and OCLC reproduce Statistical Analysis Sys- tem or Statistical Package for Social Sci- en<;:es printouts-a very good source for more details. The Library of Congress re- port reproduces, as appendix 9, the ques- tionnaires with percentages of responses received for each item, both for the Library of Congress and for respondents in the ag- gregate. All this information should prove invaluable to future planners of online cat- alogs. Furthermore, data from the transac- tion logs, as published by the University November 1983 of California and OCLC' s volume 1 should be helpful in showing (to take one example) what the users of various li- braries had done just before pressing the HELP key. Many libraries will wish to ob- . tain the computer tapes for questionnaire data, available at $50. The investigators were right in limiting the survey to library clientele, and so avoiding unmanageable complexity. However, it should not be forgotten that needs of library staff are important also. Nor should it be assumed that a popular vote is the sole test of effectiveness. There may be small minorities whose work is so- cially significant and who badly need some features that the vast majority ig- nore or perhaps find confusing. Before stakes are driven down too far, these mat- ters also should be investigated. There is still much to learn, many ques- tions yet to be answered, and some to be asked, nevertheless this massive project has advanced our understanding consid- erably. REFERENCES 1. Research Libraries Group, Inc., Public Online Catalogs and Researc_h Libraries (Stanford, Calif.: The Corporation, 1982); Rosemary Anderson and others, Library of Congress Online Public Catalog Users Suroey: A Report to the Council on Library Resources (Washington: Library of Congress, Office of Planning and Development, 1982); Joseph R. Matthews, A Study of Six Online Public Access Cata- logs: A Review of Findings (Grass Valley, Calif.: J. Matthews and Associates, Inc., 1982); University of California, Division of Library Automation and Library Research and Analysis Group, Users Look at Online Catalogs: Results of a National Suroey of Users and Non-users of Online Public Access Cata- logs (Berkeley, Calif.: Office of the Assistant Vice President-Library Plans and Policies, University of California Systemwide Administration, 1982) . 2. John E. Tolle and others, Current Utilization of Online Catalogs: Transaction Log Analysis (Dublin, Ohio: OCLC, 1983); Karen Markey, Online Catalog Use: Results of Suroeys and Focus Group Interoiews in Several Libraries (Dublin, Ohio: OCLC, 1983); Neal K. Kaske and others, A Comprehensive Study of Online Public Access Catalogs: An Overoiew and Application of Findings (Dublin, Ohio: OCLC, 1983). Other documents closely related to the proje.ct include the following: David C. Miller and Daphna Baratz, Data Collection Manual and Sampling Plan (Stanford, Calif.: The Research Libraries Group, 1982); Ray R. Larson, Users Look at Online Catalogs. Part 2: Interacting with Online Catalogs (Berkeley, Calif.: Division of Library Automation and Library Studies and Research Division, Office of the Assistant Vice President-Library Plans and Policies, University of California Systemwide Ad- ministration, 1983); University of California, Division of Library Automation and Library Studies and Research Division, University of California Users Look at MEL VYL: Results of a Suroey of Users of th~ University of California Prototype Online Union Catalog (Berkeley, Calif. : Division of Library Auto- mation and Library Studies and Research Division, Office of the Assistant Vice President-Library Plans and Policies, University of California Systemwide Administration, 1983); Council on Library Resources, Inc., Online Catalogs: Requirements, Characteristics and Costs . Report of a conference sponsored by the Council at the Aspen Institute, Wye Plantation, Queenstown, Maryland, De- cember 14-16, 1982. Comp. and ed. by Davis B. McCarn. (Washington: The Corporation, 1983). Online Catalogs 467 3. University of California, Users Look at Online Catalogs, p.3. 4. Council on Library Resources, Online Catalogs, p.108, referring to the OCLC report . 5. Research Libraries Group, Public Online Catalogs, p.35-36. 6. University of California, Users Look at Online Catalogs, p .80. 7. Research Libraries Group, Public Online Catalogs, p .24. 8. Miller and Baratz, Data Collection Manual, p .16. 9. Ibid ., p.2. 10. Matthews, A Study of Six, p.vii. 11. Kaske, Comprehensive Study, p.15-16. 12. Council on Library Resources, Online Catalogs, p.63-64. 13. University of California, University of California Users Look at MEL VYL, p .52. 14. Larson, Users Look at Online Catalogs, p.38. 15. Research Libraries Group, Public Online Catalogs, p.21. 16. Miller and Baratz, Data Collection Manual, p.13. 17. Research Libraries Group, Public Online Catalogs, p.58-65. 18. University of California, University of California Users Look at MEL VYL, p.28-29. 19. University of California, Users Look at Online Catalogs, appendix G, p.24. 20. Markey, Online Catalog Use, p.61. 21. University of California, Users Look at Online Catalogs, p.28 . 22. University of California, University of California Users Look at MELVYL, p.66-67. 23. University of California, Users Look at Online Catalogs, p.34-39 . 24. Ibid., p .107; Anderson, Library of Congress, appendix 9. 25. University of California, Users Look at Online Catalogs, p.107-10. 26. Markey, Online Catalog Use, p.90. 27. University of California, University of California Users Look at MEL VYL. p.86-87. 28. Based on: University of California, Us~rs Look at Online Catalogs, p.5,10. 29. Ibid., p.50-51. 30. Matthews, A Study of Six, p.95. 31. Kaske, Comprehensive Study, p .28. 32. Alan Seal, "Experiments with Full and Short Entry Catalogues; A Study of Library Needs," Li- brary Resources & Technical Services 27:144-55 (Apr./June 1983). 33. Council on Library Resources, Online Catalogs, p.35. 34. University of California, University of California Users Look at MEL VYL, p .x . 35. Ibid., p.77. 36. Miller and Baratz, Data Collection Manual, p.12-21. STATEMENT OF OWNERSHIP AND MANAGEMENT College & Research Libraries is published six times a year by the American Library Association, 50 E. Huron St ., Chicago, IL 60611 . Annual subscription price, $17.50 . American Librarx Association, owner; C. james Schmidt, editor. Second-class postage paid at Chicago, Illinois . Printed in U .S .A . As a nonprofit organization authorized to mail at s pecial rates (Section 132.122, JTostal Service Manual), the purposes, function, and nonprofit status of this organization, and the exempt status for federal income tax purposes, have not changed during the preceding twelve months. EXTENT AND NATURE OF CIRCULATION (" Average" figures d eno te the number of copies printed each issue during the preceding twelve months; "Actual" figures denote number of copies of single issue published nearest to filing date-the July 1983 issue .) Total number of copies prrnted: Average, 12,392; Actual, 11,836. Paid circulation: not applicable (i.e., no sales through dealers and carriers, street vendors and counter sales.) Mail subscriptions: Average, 10,688; Actual, 10,319. Total paid circulation: Average, 10,688; Actual, 10,319 . Free distribution by mail, carrier, or other means, samples, complimentary and other free copies: Average, 1,376; Actual, 1, 370. Total distribu- tion : Average, 12,064; Actual, 11,689 . Copies not distributed: office use, left-over, unaccountea, spoiled after printing: Average, 328; Actual, 147. Returns from news agents: not applicable . Total (sum previous three entries): Average, 12,j92; Actual, 11,836. Statement of Ownership, Management, and Circulation (PS form 3526, July 1982) for 1983 filed with the United States Post Office, Postmaster in Chicago, Illinois, September 30, 1983. Disco"er ll1e 1l1i1~l \~orlcl A fascinating collection on microfilm of 19th- century and early 20th-century life in third-world regions recorded in hundreds of volumes of letters from American missionaries. Papers of the American Board of Commissioners for For- eign Missions is one of the most significant collections of corres- pondence dealing with life in third-world regions from 1831 to 1919. One of the most heavily used archives at Harvard Universi- ty's Houghton Library, this microfilm collection is a primary source of research for anyone studying life in Africa, Asia, India, Ceylon, the Near East, China, the Pacific Islands, and the Americas during this time period. The collection, priced at $26,900, is available in 6 units. The first 3 units are now available. Unit 1 -Letters from the Office of the Board to $5,250. Foreign Correspondents, 1834-1919, 148 reels Unit 2- Letters from Missions in Africa, 2,250. 1834-1919, 75 reels Unit 3- Letters from Missions in Asia, 1831-1919 4,575. Prices slightly higher outside U.S. and Canada. To place an order, or for more information, call or write: 12 Lunar Drive/Drawer AB Outside No. and So. America: Woodbridge, CT 06525 m P.O. Box 45 (203) 397-2600 Reading, RG1 8HF England TWX: 710-465-6345 TEL: 0734-583247 FAX: 203-397-3893 TELEX: 848336 NADL G Research . Publications r------------------------~ : 'Mail Call so: 1 __ Please forward more information on Papers of the American 1 1 Board of Commissioners for Foreign Missions. 1 1 __ I wish to place an order. Please contact me. 1 1 Name ______________________________________ ___ Title I : Institution : 1 Address 1 I City State Zip I Phone ______________________________________ _ ~------------~-----------~