College and Research Libraries ALBERT H. RUBENSTEIN, DAVID J. WERNER, GUSTAVE RATH, JOHN A. KERNAGHAN, ROBERT D. O'KEEFE Search Versus Experiment- the Role of the Research Librarian Medical researchers, clinical specialists, and their supervisors covered by this survey, did not make frequent use of the services of research librarians, even after these services were specifically introduced to them. The potential value of a research librarian seems to depend upon the early education of the researcher and his awareness of the research librarian~ s professional capabilities. INTRODUCTION DESPITE SUBSTANTIAL ADVANCES RECENT- LY MADE in the indexing, abstracting, storing, and transmitting of scientific and technical information, many re- searchers continue to duplicate experi- ments rather than attempt to retrieve the needed information. They find that duplication is often less time-consum- ing and more certain in its results than is searching through various informa- tion channels with no guarantee of suc- cess. Yet, most researchers will also ad- mit that although duplication often serves a real function, constant repeti- tion of routine work wastes valuable re- sources and decreases real output. New information systems aim to re- duce this duplication by making retriev- al the more attractive alternative. A multiplicity of highly specialized, mis- The authors are associated with the De- partment of Industrial Engineering & Man- agement Sciences~ The Technological In- stitute~ Northwestern University, Evanston, Illinois. Work in this area has been support- ed at Northwestern since 1966 by the N a- tional Library of Medicine, National Insti- tutes of Health. 280/ sian-oriented services, however, do al- ready exist in thousands of special li- braries in universities, research insti- tutes, and corporations. Often these ser- vices are by-passed by researchers either because they are ignorant of their avail- ability or because they think services will be too slow and cumbersome. While we experiment with new infor- mation retrieval systems, we should also inform researchers about current ser- vices and encourage their use. The ser- vices of research librarians, in particu- lar, of medical librarians, best illustrate this issue. METHODOLOGY Sites and Participants As the research was designed to study group reactions to the introduction of a new information service, only well- defined work groups were selected. Be- cause specific organizational constraints and specific disciplines often affect in- formation-seeking behavior, groups were chosen from several hospitals and from various disciplines as well as vari- ous orientations (clinical, research, or supervisory). (A "clinician," as used in this study, spends at least 75 percent of his time in patient care; a ''researcher" • spends at least 75 percent of his time in pure or applied biomedical research; and a "supervisor" administers either patient care or various research activ- ities. 0 ) In the Chicago area, the specialties ·of oncology and cardiology yielded a sam- ple that satisfied the above criteria. Six hospitals were selected, with both an on- cology and a cardiology group from each. Membership in the 12 groups ini- tially numbered 110. Eleven individuals were not able to participate beyond the first stage of the experiment. Of the 99 individuals who continued to partici- pate, 71 held M.D.s, 11 held Ph.D.s, and 4 held both degrees. Five of the groups, three in cardiology and two in oncology, reported a high research orientation. Four groups reported a high clinical orientation (two in each discipline), and the remaining three groups report- ed a dual research-clinical orientation. Phases of the Field Experiment The research program was divided into three phases. ( 1) A detailed ques- tionnaire was administered to each par- ticipant to gather data about his person- al background, educational experience, organization environment, and present habits of information-seeking. ( 2) An experimental information retrieval sys- tem was introduced into each group's ':vork environment, and the participant's reaction to it was carefully monitored. Each group had access to the system for seven weeks. The retrieval system linked participants with a remote "storehouse" of information-the John Crerar Li- brary in Chicago-both by means of a 0 The participants in this study performed four types of work: research, patient care, ad- ministration, and teaching. For the purposes of this study, the time each participant spent teaching his specialty was eliminated from con- sideration. Three categories were thus formed: researchers, clinicians, and supervisors. With teaching eliminated, no participant spent less than 75 percent of the remainder of the time working in his own area of specialization. Search Versus Experiment I 281 telephone for placing requests and a facsimile system for the immediate transmission of requested material . to the participant's place of work. A pro- fessional medical librarian was the fo- cal point of the system. She received all requests for information and per- formed the necessary searches. ( 3) Six months after the service was discontin- ued, each participant was asked to eval~ uate the system, and his evaluation was compared with his actual use of the sys- tem. · REsULTS AND DiscussiON A Case History of Search Versus Experiment During the pilot study for the field experiment, ·an event occurred that put into sharp focus the relative merits of search versus experiment. 1 A researcher, working in the basic research depart- ment of a major Chicago area hospital, reported that he intended to repeat an experiment because he did not think "it would be worth the time and the trou- ble" to attempt to retrieve the informa- tion he needed. The information he needed was to answer the question: "Can substance X be successfully treated by process Y?" He knew that other less refined sub- stances had been treated with this proc- ess, but he was not sure if exactly the same process would work with such a highly refined material as X. As he did not know where the information was readily available, he repeated the experi- ment. The experiment required four hours of a technician's time, spread over two days because of a required waiting period. The researcher spent approxi- mately twenty minutes to explain the experiment to the technician. If Dr. B had chosen to retrieve the information, he would have had to perform the search himself, with little guarantee of success, because the data was not avail- able in the sources he knew. He thus de- 282 I College & Research Libraries • July 1973 cided not to risk his time on what might have been a lengthy task. A few days after Dr. B had made his decision, Mrs. A -a specially trained medical librarian-was assigned to work full-time in Dr. B's program. The li- brarian was given all the facts from Dr. B regarding his question, including information about the other applica- tions of the process he wished to use. She was instructed to keep an accurate account of the time she spent and the sources she used in finding the informa- tion Dr. B needed. Her results are pre- sented in Table 1. The ti1ne taken by Mrs. A to find the necessary information was thirty-six minutes. She spent 70 percent of this time checking written sources that proved to be of no assistance. Because she happened to be at the Crerar Li- brary when she received the request, she decided first to check the written sources. She indicated, however, that if she had had immediate access to a tele- phone at the outset of her search, she would have tried the verbal sources first. She was relatively sure that she could obtain the information through person- al contacts. Had she initially used the verbal channel, it would have taken her only 11 n1inutes to retrieve the infor- mation for Dr. B. (The brevity of her search did hinge upon having a number of personal contacts, but that is not un- usual. Part of a special librarian's job is to cultivate such sources of informa- tion.) Although it did take Mrs. A thirty-six minutes to find the information, that was far less an investment of time than the four hours it took Dr. B and his technician to achieve the same results. Although Dr. B used only twenty min- utes of his own personal time, had he been able to delegate the search to Mrs. A, it might have taken him less than half that time to explain to her what he needed. If the research librari- an had been a familiar source for Dr. B, he probably would have consid- ered a search the more attractive alter- native than a repetition of the experi- ment. Searching would have been less costly and time consuming than dupli- cation. Information-Seeking Patterns This case history is by no means unique. During the first phase of the TABLE 1 TIME AND RESULTS OF THE LIBRARIAN's INFORMATION SEARCH Source Used John Crerar Library card catalog 0 Open shelves at Crerar Library Telephoned a friend at a chemical company who was familiar with processes related to Y Telephoned Dr. C but spoke to a Dr. D Telephoned Mr. P at the pharma- ceutical house TOTAL Time (in minutes) Results 10 Negative 15 Negative 5 Referred to a Dr. C at another hospital. 3 Referred to another doctor at the same hospital, who recommended Mr. P at a pharmaceutical house. 3 Yes, X can be treated by process Y. He 36 gave the names of two pharmaceutical houses that could supply the requested information. 0 The John Crerar Library is a privately supported public library located in Chicago and serv- ing science, engineering, and medicine. Its collections contain over 1,000,000 volumes. J _I 1 j j j • full field experiment, the participants were asked to ( 1) "indicate with a check mark how frequently you use each of the listed information sources" and ( 2) to report "what have been the main sources of influence in the development of the way in which you currently keep yourself informed." The results of the researchers' and clinicians' reported fre- quency of use of the listed information sources are presented in Table 2. Seven- ty percent of the researchers and 7 4 per- cent of the clinicians use the services of a librarian less than once a month. In comparison with the general pat- terns that emerge from Table 2, the low utilization of the librarian is sur- prising for three reasons. First, since both groups make relatively frequent use of written sources and libraries, on·e would expect a comparably frequent use of librarians. However, this is not the case. Second, the researchers' use of written sources and libraries is signifi- cantly greater than that of the clini- cians, one would expect an equally greater utilization of librarians. There Search Versus Experiment I 283 is, however, no significant difference be- tween the researchers and the clinicians in the frequency with which they use librarians. Third, although both groups reported a high verbal orientation- de- pendence on colleagues, subordinates, and supervisors for information, the li- brarian is utilized much less than other verbal sources of information, even though his value as a verbal source can often be greater (as our case history il- lustrated) than an individual's other personal contacts. It seems that an indi- vidual's use of a research librarian is not directly related to the frequency with which he uses a library; and that neither the researchers nor the clinicians regard the librarian as a valuable col- league. This confidence gap between scientists and librarians has been noted by other studies. Slater and Fisher, during a study of the use made of technical libraries in Great Britain, observed that academic scientists were unwilling to delegate lit- erature searches to the librarians in their respective universities. 2 A 1969 re- TABLE 2 FREQUENC Y OF SOURCE USE OF SCIENTIFIC AND TECHNICAL INFORMATION Sources of Scientific Researchers ( N-46) Clinicians ( N-42) ;.:.:..., and Technical Information 1 2 3 1 2 :l X2 WRITTEN: Textbooks 25% 40% 35% 28% 63% 9% 9.65~n Scientific Books 21 % 36% 43% 33% 46% 22% 5.09° Handbooks 17% 51% 32% 48% 36% 16% 10.40° 0 Professional Journals 4% 41% 55% 16% 28% 56% 5.21 # Technical Publication s 25% 51% 24% 68% 19% 13% 18.35° 0 VERBAL: Colleagues 12% 34% 54% 6% 31% 63% NS Subordinates 39% 41% 20% 22% 42% 36% NS Superiors 25% 54% 21 % 29% 40% 31% NS Librarians 0 70% 24% 6% 74% 26% 0% NS COMBINATIONS: Libraries 20% 54% 26% 36% 54% 10% 6.23° 0 1-Less Than Once a Month op .10 2-More Than Once a Month But oop .05 Less Than Four Times a Week NS-Not Significant 3-More Than Four Times a Week 0 The librarians referred to in this study were all associated with medical libraries-such as a general medical library like the Crerar Library, or a specially oriented research library like that of the Chicago Argonne Cancer Research Hospital. 284 I College & Research Libraries • July 1973 port from the British Office for Scien- tific and Technical Information re- vealed that, among a sample of students engaged in chemical research, more than 52 percent had never asked a member of the library staff for help in perform- ing literature searches, while 39 percent asked for such help only on rare occa- sions.3 Hall has observed that "despite their apparent academic status, librari- ans are seldom regarded as an equal by their colleagues in teaching and re- search." He outlines factors possibly ac- counting for this situation, which will be discussed later. 4 Researchers in both pure and clinical sciences often lament the vast scattering of important information in books and journals, making it impossible to keep track of what is being published. If a researcher is unable to read a specific journal and asks a colleague for a de- scription of what it contains, at best he will receive a summary description com- mitted to an often faulty memory. A research librarian, on the other hand, has immediate access to all the written sources and should be able to provide a full and accurate account. Such a li- brarian can also regularly supply cur- rent tables of contents, bibliographies of new dissertations, and reprints of new articles according to subject area. The services of a trained librarian could be one of the researcher's most important means for keeping abreast of what is being published in his own field and in related specialties. If these re- searchers and clinicians were aware that a research librarian like Mrs. A could save them a great deal of time and mon- ey, the frequency With which they would use his or her services might be quite different. One reason many researchers and cli- nicians are unaware of the potential value of a research librarian (and thus do not regard him or her as a profes- sional colleague) can be traced to their past educational training in the use of information sources.5 When the partici- pants in this experiment were asked to indicate which factors had been most influential in determining their present style of seeking information, only 13 percent of the researchers and 2 percent of the clinicians mentioned specific edu- cation in the use of information sources. (See Table 3. ) This lack of training can be account- ed for, in part, by the lack of activity on the part of library directors and li- brarians. The results of a national sur- vey of the user services offered by med- ical school libraries in the United States revealed that only twenty-one of the ninety-two libraries surveyed provided "formal" instruction ( defined as re- quired or elective courses) designed to help medical students and faculty make optimal use of information sources and services. 6 The authors go on to com- ment that ~~at over half of these twenty- one libraries, the courses rely solely on passive instructional methods-lectures, assigned readings, etc.-as contrasted to methods that require the active partici- pation of students, such as small group seminars, practical exercises, etc." In most of the libraries that offer only ~~in­ formal'' instruction, such instruction usually consists of a one- or two-hour orientation tour or lecture. The authors conclude that: instruction is an area where there is great room for improvement. The usual orientation tour or lecture ac- complishes little that cannot be better done by a good written guide to the library and its services; and passive in- structional modes are inadequate for teaching the optimal use of informa- tion resources. Lack of financial resources or the ab- sence of enough trained personnel, may prevent many libraries from offering de- tailed and on-going instruction in the ·use of information resources. However, if such instruction is to be offered, it seems that the library must be the focal A ., point for training. Our own informal inspection of the courses offered by six major medical schools in the midwest- ern and western United States revealed that only . one school listed a course in the use of information services, and that course concerned computer use for TABLE 3 MosT IMPORTANT INFLUENCE IN FoRMING RESEARCHERS' AND CLINICIANS' INFORMATION- SEEKING STYLE Researchers Clinicians Influence (N-46) (N-42) Personal (colleagues) 33% 50% Written Sources 31% 17% General Education 23% 31% Specific Education in the Use of Information Services 13% 2% 100% 100% the storage and retrieval of informa- tion. Although universities should be en- couraged to offer more courses in the use of information sources and services, the responsibility for training students and faculty in the use of library services can only rest with the library staff itself. The Participant's ·Use of the Experimental Information System During the second phase of the ex- periment, the participants were intro- duced to a new (to them) information retrieval ~ystem. A professional medical librarian was stationed at the John Crerar Library. The participants could request any type of information, rang- ing from requests for reprints of spe- cific articles, literature searches, and an- swers to specific questions such as the one posed by Dr. B in our case history. They could request this information by phone or letter and once the search was performed, the information was sent to them immediately by means of a fac- simile system connection between the li- brary and each group's work area in the various hospitals. Each group had access to the system for seven weeks. Search Versus Experiment I 285 . Of the ninety-nine individuals who were exposed to the new system (forty- six researchers, thirty-one clinicians, 0 0 and twenty-two supervisors), forty in- dividuals used the system to make a to- tal of 141 requests. The percentage of use by each category is shown in Table 4. TABLE 4 UsE OF THE NEw INFORMATION SERVICE BY EACH CATEGORY Users Nonusers Category (N-40) (N-59) Researchers ( N -46 ) 52% 48% Supervisors ( N-22) 36% 64% Clinicians ( N-31) 23% 77% TOTAL (N-99) 40% 60% The significance of a 40 percent use of the new information system is de- batable. However, in our context, the most interesting part of the data con- cerns the type of requests made of the librarian stationed at the John Crerar Library. Of the 141 requests for infor .. mation, 85 were for copies of specific articles, 50 requests were for either re- prints of existing bibliographies or the compilation of new bibliographies in a given subject area, and only 6 requests were for literature searches or answers to specific questions. The behavior of the researchers and clinicians, in terms of the types of requests they made of the librarian, corroborate the previously discussed ccconfidence gap" between re- searchers and librarians. CoNCLUSIONs Although · great emphasis is currently on developing complex new systems for information retrieval, quite possibly a greater emphasis needs to be placed :upon informing users about current in- formation services. encouraging them to use these services, and making the ser- ()O Eleven of the clinicians did not partici- pate in this phase of the study. - ---·----------------------------- -- 286 I College & Research Libraries • July 1973 vices more responsive to their individu- al needs. Librarians can try to establish themselves as professional colleagues in the eyes of the researchers with whom they interact. As Hall observed, librari- ans must begin by "participating more actively in the affairs of the university and by insuring that they are always de- ployed on work which is commensurate with their abilities."7 This means that librarians should "advertise their ex- pertise in information retrieval." He goes on to remark that one "way of breaking down the physical and psycho- logical barriers between library and teaching staffs ... is to establish a li- brarian/information officer" who will be directly responsible for one specific subject area and who will actively make contacts in each university department or research institute that deals with that subject area. Through this type of inter- action, the teaching or research staff will learn that the librarian "has some knowledge of their subject area and will be free to concentrate on the practical benefits of such a service." Although it is extremely difficult to alter the information-seeking style of mature researchers, the librarian can try to reach the younger researchers, the un- dergraduate and graduate students. The preliminary results from a planned five- year longitudinal study of how infor- mation-seeking style is formed among medical students in the United States in- dicates that these students are not well- trained in the use of information sources and do not take advantage of all the information resources available to them. 8 If researchers could be trained in the efficient use of informa- tion systems and services, existing sys- tems and services could function more effectively. REFERENCES 1. David J. Werner, "A Study of the Relation- ships between Some Task, Professional, Per- sonal, and Organizational Characteristics and the Use of an Experimentally Intro- duced Information System in a Medical En- vironment," a Ph.D. dissertation, Depart- ment of Industrial Engineering and Manage- ment Sciences, Northwestern University, Document Number 69/25, June 1969. 2. M. Slater and P. Fisher, "Use of Technical Libraries," ASLIB Occasional Publication no.2, London, 1969. 3. Office for Scientific and Technical Informa- tion, "Students Chemical Information Proj- ect," Final Report: Part 1, February 1969; Part 2, June 1969, London. Of course, in some universities, students are not encour- aged to request help in actually doing a lit- erature search, although they are generally encouraged to seek guidance in their search strategy. 4. J. Hall, "Information Services in University Libraries," ASLIB Proceedings, v.24, no.5 (May 1972), p.293-302. 5. Albert H. Rubenstein, "A Longitudinal Study of the Development of Information Style," in Management Information Sys- tems, ed. by Erwin Grochla. Wiesbaden: Betriebswirtschaftlicher Verlag, 1971. 6. Richard Orr, Harold Bloomquist, Gwendo- lyn S. Cruzat, and Arthur P. Schiess, "User Services Offered by Medical School Librar- ies in 1968: Results of a National Survey,'' Bulletin of the Medical Library Association 58, no.4 (Oct. 1970), p. 455-92. 7. J. Hall, "Information Services ... ," p.301. 8. Albert H. Rubenstein, "The National Insti- tutes of Health Longitudinal Study Progress Report: The Development of Information Style by Researchers," Department of Indus- trial Engineering and Management Sciences, Northwestern University, Document Num- ber 71/46, August 1971. See also, this au- thor and Gustave J. Rath, Robert D. O'Keefe, John A. Kernaghan, Evelyn A. Moore, William C. Moor, David J. Werner, "Behavioral Factors Influencing the Adop- tion of an Experimental Information System by Medical Researchers and Clinicians in Six Hospitals," Catalog of Selected Docu- ments in Psychology: 2:101-2 (Summer 1972). t