College and Research Libraries JOHN A. KERNAGHAN, SALVINIJA G. KERNAGHAN, ROBERT D. O'KEEFE, and ALBERT H. RUBENSTEIN The Influences of Traditional Services on Library Use This study, on the relationship between user preferences, library charac- teristics, and the fr eq u ency of library use, employing a sample of 655 students in fi ve medical schools, supports in a tentative fashion some tra- ditional principles relating library characteristics to frequency of use. Other factors assumed to influence library use, such as the availability of au- diovisual materials or th e physical comfort of the library , do not appear to be related to frequency of use. THE SUCCESSFUL PLANNING AND DESIGN of any information system must ultimately de- pend on an understanding of user needs and preferences in seeking information. One important measure of a library's success as an information system is the degree to which its various services and facilities are used by its potential user group. This mea- sure is simultaneously a good indicator of how well the library 's planners and operators understand user needs related to information seeking. A recognition of this relationship between users' needs and preferences in information john A. Kernaghan is assistant professor of administrative studies, Graduate School of Busi- ness, Indiana University, Bloomington . Salvi nija G. Kernaghan has been f eature editor of Hospi- tals and J .A.H.A. and senior associate editor of Respiratory Care . Robert D . O'Keef e is associate professor of marketing, School of Business , De- Paul University, Chicago . Albert H. Rubenstein is professor and direc tor of th e Program of R e- search on the Management of Research and De- velopment , Department of Industrial Engineering and Management Science, Technological Institute, Northwestern University, Evanston, Illinois. The resea rch reported here has been supported by the National Library of Medicine, National Institutes of H ealth , Public Health Service , PHS Grant #LM00098 . A portion of this article was pre- sented at the national meeting of the Institute of Management Science , November 1976. 214 I seeking and actual library use has long been a subject in library literature. Indeed, we might consider articles in which limited use of a library is attributed to some "fault" in the library information system as a separate "diagnostic-prescriptive" category in library literature. For example, in attempting to relate stu- dents' library use to their academic achievement, Ritter found no correlation between library book borrowing and grade point average. 1 Ritter's prescription for in- creased use puts the burden on " the classroom . . . as of supreme importance for motivating the use of the library whether by way of formal requirements or by way of subject interest." Lubans makes a similar recommendation, after studying the per- centages of nonusers in a technical- engineering course of study. 2 Other prescriptions include cenh:alizing collections for students' "convenience" and providing users with "those books that have a frequency of use that is so low, or a cost. that is so high , as to prohibit individual purchase . "3 Rzasa and Moriarty offer a simple formula for increasing library use: Decide what kind of reader group you wish to attract, increase the portion of the collection that will inter- est that group, and let the group know the books are available. 4 Martin and his coauthors repeat the fre- quently heard but seldom heeded advice that librarians offer courses in methods of information seeking. 5 Articles like these share the assumption that, if a library will only institute such- and-such a change, it will foster increased and more efficient use of its services. The operative word here is assumption. Al- though much of their advice may indeed ef- fect the desired results if implemented, the authors of these articles offer no empirical support to encourage librarians to embrace such "improvements" and to similarly per- suade administrators and budget committees. The library field has yet to assemble con- vincing evidence that the traditional "prin- ciples" upon which libraries continue to be planned and operated have a direct relation to user needs. Certainly, we know that some percentage of each library's potential user group uses some of its services. What we do not know-although the literature makes frequent guesses-is why libraries are not used more often and more efficiently and whether we can really expect a suggested improvement to make a dif- ference. PURPOSE AND THEORY The purpose of the study reported here is to begin to investigate the assumed cause- effect relationship between these traditional principles and frequency of library use. It is part of a larger study in which we have ex- amined the impact of various characteristics of the information environm e nt on the de- velopment of the information-seeking styles of medical students in a number of U .S. medical schools. 6 It is reasonable to theorize that a person ' s preferences in acquiring needed information -that is , the pattern or style of information seeking-have not developed at random . Indeed, we suggest that a person's style of searching for information is a direct result not only of the kind of information needed; it is also influenced by current and past constraints on seeking information and by that person's previous experience with (i.e., success or failure in using) various sources and information-seeking methods. . Because the library is considered a major information resource in any academic or professional training environment, a portion Traditional Services I 215 of the larger study was devoted to examin- ing the library's significance in the patterns of information seeking that were reported by the sample .of medical students. Specifi- cally, the library study reported here was designed to test the general notion that the more a library implements the principle of providing service to all potential users, the more frequently it will be used. This prin- ciple was expressed in the form of the fol- lowing propositions (in which may also be found the " prescriptions" that often appear in the literature for improving library utili- zation rates): 1. The more extensive a library's collec- tion, the more frequently the library will be used. 2. The more a library makes audiovisual materials available to its users, the more frequently the library will be used . 3. The more effective a library's recall policy for the timely return of materials from all borrowers, the more frequently the library will be used. 4. The more a library makes recent and current materials available to its users, the more frequently it will be used. 5. The more assistance a library staff pro- vides to users in their search for information sources, the more frequently the library will be used. 6. The more comfortable the physical en- vironment of the library, the more fre- quently it will be used . 7. The more user seating a library pro- vides , the more a library will be used as a reading/study room. 8. The more a library is used as a reading/study site, the more it will be used for other services. DATA COLLECfiON Five major medical schools in the Mid- west and on the West Coast were chosen as sites for collecting the student sample data as well as the information on the medical li- braries that the students used. School D was the site of the smallest sample group- 120; it was followed by school E with 125, school C with 176, school B with 266, and school A with 314 . On-site associates of the research project administered a written questionnaire to all students in each school who were willing to 216 I College & Research Libraries • May 1979 participate ; the sample included students from the first through the fourth year of training. (Total usable responses were 655 out of 1,001, or a response rate of 65 per- cent.) In order to describe their information- seeking patterns, students were asked to re- · spund to a number of questions that simply required checking off appropriate charac- teristics from a list of those that might de- scribe their habits , their sources, and other factors relevant to their information seeking. A few other questions asked for responses in other forms. To indicate how often they used specific information sources, for exam- ple, students were asked to check the col- umn that represented the appropriate number of uses per period . Some questions asked that sources be given numerical ranks according to their importance to students, and others asked that they briefly describe their typical approach to seeking informa- tion to solve a specified problem. We chose to have students report on their own use of sources rather than to de- pend on records of use of sources (e . g., books checked out) because the latter ap- proach would have described students' use of only a few sources of information . The point of the entire study was to arrive at as comprehensive a picture as possible of stu- dents' preferences in using all the sources available to them, including those for which no records would be kept, such as their using their personal libraries and their con- sulting their professors. Since the library study was only a portion of the larger study, we gathered all data on student use patterns-whether of libraries or other sources-in a uniform fashion , through stu- dent self-reports. As was mentioned above , the library sample that is represented in this portion of the study consists of the five medical school libraries that served the student sample. (Some hospital libraries in institutions where students were assigned for clinical training also participated in the study; data from these sources will be analyzed in a later report.) Information was collected from the medi- cal school libraries by way of a mailed, self- administered questionnaire that was com- pleted by the director of library services or some other designated person. The ques- tionnaire consisted of two sets of identical questions-one in which information about services during an earlier period was re- quested, the other in which current infor- mation was to be provided. The purpose of collecting retrospective data was to identify any significant changes in library services during the period in which data were being collected from the student sample. How- ever, almost no changes were reported, and the information regarding libraries used in this analysis all comes from the current data portions of each questionnaire. The questionnaire items were developed to provide data on each indicator in the three categories of information according to which the libraries would be compared- holdings, staff service to users , and library physical characteristics. The result was an instrument that could provide us with a comprehensive summary description of each medical library; only those portions of the data appropriate for comparison with stu- dent data have been used in this report. The concept of designing most questions in multiple parts, so that each consecutive part represented a greater degree of service, was modeled on the approach used by Orr and his colleagues in an earlier study . 7 Unlike our instrument, however , their question- naire was administered by an interviewer. An example of this kind of multilevel question is the request that the librarian de- scribe (simply by checking the appropriate box or boxes) the kinds of services that the library staff will provide to the user who needs some references on a particular topic. The librarian may check one, a few, or all of the following : (a) will provide guidance in compiling a list; (b) wlll find a reference list already compiled elsewhere ; (c) will do a quick subject search and thus provide some references; and (d) will compile a compre- hensive list. In addition, such questions describing service to users asked that each user group that benefits from the service be identified, again by checking the appropriate box. In this way , services . that were provided to some users (physicians and/or faculty) and not to others (students, interns , and resi- dents) were distinguished from those ser- vices that were equally available to all users. DATA ANALYSIS Schools were ranked in descending order according to positive or favorable student comments for each of the variables. The data for each student variable were submit- ted to a standard chi-square test. Because there were only five data points in the medical school library sample, these data were not amenable to statistical analysis . Ranking of libraries was simply ac- complished by assignment of points for ser- vices provided, according to the libraries' responses to the questionnaire . The sum of points on a particular variable (detailed examples appear in the discussion of results) was the basis for ranking libraries according to their own reports of service and such other factors as the library's physical charac- teristics. There was only one exception made to this ranking scheme; this appears in the following discussion of libraries' hold- ings, in which libraries are compared on ac- tual number of items in three categories of their collections and then ranked accord- ingly . The final step in the data analysis con- sisted of comparing the order of rank gener- ated by student responses to a particular variable with the order of rank based on each library' s reports regarding that vari- able. For example , libraries were ranked according to students ' perception of the availability of an adequate audiovisual col- lection in their library. Libraries were also ranked according to their own description of the audiovisual collection they provided to various groups of users. These two rankings-the one based on student perceptions and the other based on library self-reports-were then compared to see if the libraries' order of rank in each was the same or comparable. Theoretically, the order of rank generated from student and library responses should be identical since they were evaluating or describing the same library characteristic. However, we found some dissimilarities and, in these cases, we attempted to suggest plausible reasons for the differences. The next step was to compare the librar- ies' ranking based on frequency of use as the students had reported it, first with the student perception ranking and then with Traditional Sen ;ices I 217 the library self-report ranking. We made these comparisons for every variable, i.e ., according to each library characteristic iden- tified in this study. By doing so, we hoped to discover whether either of the two rank- ings based on descriptions of library service-or neither of them or both-might be good predictors of frequency of library use. For example, if the ranking based on student perceptions of a particular library characteristic seemed more in line with the frequency of use rank than was the library self-report ranking, we again speculated about the possible reasons for this relation- ship. RESULTS To arrive at a ranking of the libraries according to frequency of use by those re- sponding, the questionnaire asked the re- spondents to indicate how often they used their school library "to obtain and work with medical information." The respondents had a choice of checking one of the following : not at all, less than once a month , one to three times a month , one to three times a week , four to five times a week, or daily. Frequency of Use The summary of responses on frequency of library use is shown in table 1. Responses for the three periods of most frequent use (e, f, and g in the table) were added for each library and divided by the total re- sponses for that library to compute a per- centage . A chi-square test indicated a dif- ference among the responses that was sig- nificant at the .005 level x2 = 64.52). In assigning ranks on this and all other student-reported measures, a tie rank was assigned when a library was within three percentage points of a library scoring more . In table 1, for example, library B is given the same rank as library E, which is only two percentage points "better" in frequency of use. Therefore, the students rank the li- braries on their frequency of use as follows: A = 5, B = 3, C = 2 , D = 1, and E = 3 . Library Holdings Because the basic service that a library can perform for its users depends on its col- lection, the libraries were ranked on their collections by comparing three indicators of 218 I College & Research Libraries • May 1979 TABLE 1 FREQ UENCY OF USE OF SCHOOL LIBRARY A Frequency of Use No. "k No. (a) No answer 5 .03 5 (b) Not at all 26 .13 4 (c) Less than once/month 44 .22 25 (d) 1-3 times/month 53 .27 59 m 1-3 times/week 49 .25 50 4-5 times/week 15 .08 9 (g) Daily 5 .03 9 Total Responses 197 1.01 161 e+f+g T .35 .42 Rank 5 3 x• = (64 .52l: df~-~;?; . each collection: (a) total number of bound volumes, (b) number of serials, and (c) aver- age number of reserve titles per semester. The information the libraries provided to this question and the consequent rankings (for each indicator as well as overall) appear in table 2. The overall rank on a measure of holdings was as follows: A = 5, B = 4, C = 1, D = 2, and E = 3 . This overall rank is based upon an average of the individual rankings on each of the three indicators, generated by permitting each of the three indicators to assume equal importance. This strategy is supported in part by the fact that each of the three individual rankings has some simi- larity to the frequency of use ranking, which suggests that each may have an independent influence on frequency of library use . The ranking on the library holdings measure was compared to the ranking for the students' frequency of use. (A ranking for students' perception of the size of library holdings was omitted because such a ques- B Librari es c D E "k No. "k No . "k No. "k .03 1 .01 3 .03 8 .06 .02 5 .06 0 9 .07 .16 8 .10 5 .05 21 .17 .37 25 .32 22 .23 37 .30 .31 33 .43 41 .43 37 .30 .06 2 .03 14 . 15 9 .07 .06 3 .04 10 .11 4 .03 1.01 77 .99 95 1.00 125 1.00 .49 .68 .40 2 1 3 tion would have provided meaningless in- formation. ) The -ranking on library holdings tended to correspond to the libraries' rank- ing on frequency of use (see table 2). These results seem to support our first proposition: The more extensive a library's collection , the more frequently it will be used . Audiovisual Materials A ranking based on libraries ' responses regarding their audiovisual material collec- tions took ·into account information about each library 's providing the following : (a) films ; (b) microfilm ; (c) audio tapes; (d ) equipment for using the above in the li- brary; (e) above A-V materials for use out- side the library; (j) equipment for using available materials outside the library; and (g) assistance in obtaining A- V materials from another source if they are not in the library's collection. To arrive at a rank, a "yes" response to each of the above indicators was assigned one ( + 1) point; a " no" response was as- TABLE 2 LIBRARY DATA FOR RANK ON HOLDINGS Indicator Libraries of 1-I oldi n~s A B c D E (a) Bound volumes 66,685 187,305 235,368 250,000 73,000 (a ) Rank 5 3 2 1 4 (b) Serials 1,808 1,300 3 ,996 2,471 2,000 (b) Rank 4 5 1 2 3 (c) Reserves/semester 300 300 2,700 800 300 (c) Rank ·3 3 1 2 3 Overall Rank 5 4 1 2 3 Frequency of Use Rank 5 3 2 3 signed minus one ( -1) ; and an answer that indicated that the library provided the ser- vice to some users (e.g., physicians and fac- ulty) but not to others (e.g., medical stu- dents) was assigned minus two ( -2) points, because such "differential service" invar- iably excluded student users from its ben- efits. The ranking of the five libraries was then based on the sum of the points assigned to all indicators for one measure. Thus, as may be seen in table 3, on the measure of availability of audiovisual materials, the li- braries are ranked as follows: A = 3, B = 2, C = 1, D = 5, and E = 3. Students' perception of the adequacy of a library's audiovisual material collection was elicited in the student questionnaire by an item asking for agreement or disagreement with the perception that the library had "in- adequate audiovisual aids." Their responses appear in table 4. Because a "yes" response indicates an unfavorable perception of a li- brary's audiovisual holding, the library with the lowest percentage of affirmative re- sponses is ranked first, and so on. The libraries are ranked as follows , based on students' perception of the adequacy of their audiovisual collections: A = 2, B = 2, C = 1, D = 5 , and E = 4. A comparison of the three rankings for this measure (based on library self-reports, on students' perception, and on frequency of use) appears at the foot of table 4. The library ranking closely matches the students' ranking based on their perception of the adequacy of audiovisual materials at their libraries. However , neither of these two rankings Traditional Services I 219 TABLE 3 LIBRARY DATA FOR RANK 0 AVAILABILITY OF AUDIOVIS UAL MATER IALS Indicators of A'