College and Research Libraries J. S. KIDD On-Line Bibliographic Services: Selected British Experiences The British recently have completed a program to test and evaluate on-line bibliographic search capabilities at six of their institutions of higher education. Some of the insights derived from the test trials are relevant to U.S. academic librarians. Outcomes include high user ap- peal and the opportunity for new or .expanded modes of service. The importance of advance planning regarding organizational Qdjust- ments, service promotion, and the specification of procedures for use of the new resources were made explicit by the test demonstrations. CoNFRONTING THE AVAILABILITY of commercial on-line bibliographic search capabilities, the academic librarian sooner or later is forced to consider the following issues: 1. Whether to subscribe and which system or combination of systems to join; 2. Where to locate the operation in both a physical and an organiza- tional sense; 3. How to promote and/ or control user demand and access; 4. What procedures-of-use to estab- lish. These are not trivial issues. The ad- vent of the capability to search computerized bibliographic files from remote stations could have subtle but ]. S. Kidd is acting dean, College of Library and Information Services, U niver- sity of Maryland, College Park. The work reported here was subsidized in part by the Research and Development Department of the British Library by means of a Con- sultancy arrangement. The author grate- fully acknowledges the support provided by Dr. P. L. Holmes and his staff associ- ates. However, the views represented are entirely those of the author and should in no way be considered as emanating from the British Library. profound effects on the ways in which college and university libraries are ad- ministered and how they are perceived by academic executives, faculty, and students. One hint as to the nature of opera- tional changes that might be induced by the presence of this relatively new mode of computer applications in libraries is manifested in the statistics which de- scribe reference activities in annual re- ports. Characteristically, the number of "'demand bibliographies" produced by the reference staff is very low .1 Insofar as formal, structured subject bibliographies are a tangible product of the reference process, they are generally produced by the user in a more or less intensively guided but primarily do-it- yourself mode. The reference depart- ment provides the tools and the guidance but usually does not formu- late the product, as such. From some preliminary observations that will be presented in detail below, it seems more than likely that the modes of use of the computer capability can lead to a greatly increased output of in- dividualized bibliographies that are for- mulated by a librarian rather than by the user (or by some other "agent," such as a graduate research assistant). /285 286 I College & Research Libraries • July 1977 One implication is that the whole conception of service productivity will change as a consequence. Other possi- bilities relate to the relative priority as- signed to the continued expansion of holdings of certain reference materials. There are also staff training implica- tions, etc. 2 These concerns have emerged in a very short period of time. An early specification was made by Gardner and colleagues in 1974.3 Their anticipation of these changes was only too valid. In brief, one can safely assert that the availability of on-line search capa- bilities, while eminently attractive, is not likely to be an unalloyed blessing. There are likely to be certain problems incumbent upon the adoption of this innovation by academic librarians. BACKGROUND It is possible that the concerns ex- pressed above, among others, prompted the responsible officials of the British Library Research and Development Di- vision ( BLRD ) to institute a series of field trials as a form of pilot test and demonstration activity early in 1974. The broad, programmatic concern ac- corded to on-line bibliographic systems on the part of BLRD and the informa- tion professions at large in Britain has had many facets. It includes a range of activities from training librarians and others in the use of on-line equipment to the complete development of domes- tic (i.e., British) systems. Important precedents were laid down in the U.S.- U.K. cooperative arrangements regarding the MEDLINE system which were re- fined in 1971 from the prior agreements established for MEDLARS in the 1960s. Thus the observations which follow are reflective of only one modest com- ponent in a complicated array of activi- ties relating to on-line bibliographic system development and evaluation. Specifically, the observations concern in- stallations in environments in which en- gineering as a subject orientation was predominant. Moreover, and congruent with these settings, the systems under examination were limited to the Lock- heed DIALOG system and the ORBIT system offered by the System Develop- ment Corporation. TRIAL INSTALLATIONS Six institutions of higher education participated in the pilot test project: the University of Manchester; the Univer- sity of Wales-University College, Car- diff; the University of Edinburgh; Loughborough University; Hatfield Poly- technic; and Cranfield Institute of Technology. In the first three listed, the installation was operated either by a computer science department or (in the case of Edinburgh) by the computer service center. In the last three, the in- stallation was operated by the library. There are other important differel).tial aspects of setting, as follows. The University of Manchester is an urban but residential institution in the industrial heartland. It is the major component in a central-city educational complex which includes the University of Manchester Institute of Science and Technology (essentially an autonomous body) and Manchester Polytechnic. This complex is geographically compact to the extent that it has been speculated that it provides one of the most dense concen- trations of university-level faculty and students in Britain. University College-Cardiff is similar except that the multi-institution "cam- pus" is more extended in space. The University of Edinburgh is ur- ban, but the density factor is moderated somewhat by the recent provision of a branch campus (Kings) in a suburban location which now contains most of the science departments. Loughborough University and Hat- field Polytechnic can be characterized as medium-sized, relatively new sub- urban facilities. Cranfield Institute is relatively small, restricted to graduate- level instruction, and self-contained in a rural environment. 0BSERV A TION AL PROCEDURES In the late spring of 1975, informal site visits were undertaken at the behest of BLRD. The objective was to provide a marginal augmentation of the main mechanisms for evaluation in the form of an "uncommitted viewpoint." For each installation, the main mech- anisms for evaluation were the records initiated and maintained by the opera- tional and managerial staff. The central but limited question to be resolved by means of on-site observations by an "outsider" was whethet: there were im- portant situational factors that could affect the outcome of the evaluation but which might be taken for granted by the operational personnel because of their intense familiarity with their own environments. In each case, the "outside observer" spent one day studying the situational factors. Interviews were conducted with the participating personnel, and a dem- onstration search was performed with the observer in the role of the user. DERIVATIVE IssuEs There were subtle situational factors that might moderate the main conclu- sions of the test trials. For example, the installation at Hatfield Polytechnic was located in the same area as an audio- visual instructional services unit. This arrangement brought the capability to the attention of a certain group of po- tential users who otherwise would have been only peripherally involved. Other issues came to light in the process of observation that seem more germane to academic librarianship in the U.S. First and foremost, it was evident - that the on-line search capability pro- vides the base for new dimensions of service that can have strong user appeal. While in no instance was there an "over- night sensation" effect, demand grew over an eight- to ten-month span at a accelerating pace based mainly on word- of-mouth endorsements by users. The "classic" user was the graduate student Bibliographical Services I 287 in the early stages of a thesis or disser- tation project. For these students and other users in a similar "project mode," the service was very welcome (i.e., "user acceptance" was high at all sites). Second, it was apparent that the li- __, brary base was superior to the computer science base. In minor part, this effect appears to result from the predilection of the computer science personnel to look upon the installation as a research tool as opposed to a service device. Per- haps more significantly, the library in- stallations could complete the service process through the stage of document delivery whereas the computer science- linked installations could not. Third, it became clear that the role _ of the "intermediary" in the service transaction was far more complicated and delicate than had been anticipated. In this context, the service transaction involves three entities: the user, the in- termediary, and the terminal. The issue can be simplistically specified as the de- termination of an optimum level of participation by the intermediary in the user-system interaction at the terminal. At one extreme, the intermediary plays a minor role in a quick initial in- troduction of user to terminal and acts thereafter as an on-demand technical consultant only when the user experi- ences some particular difficulty. At the other extreme, the intermediary acts as a user surrogate, with the user in the background or even absent during search operations. The first extreme is very inefficient, particularly for novice users. Since most users have only highly intermittent needs for demand bibliog- raphies, novice status is not easily sur- passed. The other extreme is ineffective in the same way that a batch-process search system is ineffective: real user in- terests are not always accurately trans- lated into productive search strategies by the intermediary. TENTATIVE RECOMMENDATIONS REGARDING ADMINISTRATIVE IssuES To return to the explicit issues set 288 I College & Research Libraries • July 1977 out at the beginning, the question be- comes that of whether the British experience provides either specific pre- scriptions or at least an approach to such prescriptions. Given the caveat that the experience in question is filtered through the perceptions of an "outsid- er," some recommendations can be made. The Decision to Subscribe The question of whether or not to subscribe can be answered affirmatively with the proviso that the most beneficial installations will be those in colleges and universities that have substantial re- search programs in the natural sciences or in engineering. Support capability for the social sciences and for profes- sional programs (e.g., education, busi- ness administration, librarianship, etc. ) remains spotty in spite of the availabili- ty of some relevant files such as Psycho- logical Abstracts and ERIC. Support capability for the general humanities area is not yet significant. 4 Even with such a proviso, the total constituency in the U.S. is in the range of 500 to 600 institutions, many of which should probably require more than one terminal. 5 It has been estimat- ed that only about thirty to forty such institutions have actually installed com- mercial system terminals in their mairi libraries at the time of this writing. 6 (It should be noted that there are many more than thirty to forty university sub- scribers to the commercial on-line ser- vices. DIALOG alone has more than 200 such subscribers.'7 The explanation of what would otherwise be unreconcilable data appears to be that most such sub- scriptions involve installations in com- puter science centers, special research units, or in branch libraries that serve a high-technology clientele.) Cost is probably a factor in restrain- ing the acceptance of commercial on-line services by academic library di- rectors, given the present austere eco- nomic climate in higher education. Realistically, however, the capital costs are marginal for most large U.S. insti- tutions, and operating costs are tied closely to the demand factor so that the usual ambiguities with respect to cost- effectiveness are minimal. The British experience is specifically instructive in this matter. While their austerity is significantly more intense than ours, 8 the librarians' response was to retain their installations even though the subsidies which facilitated the test and demonstration program were about to be terminated. The head librarians interviewed expressed the intent to . con- front their respective academic execu- tives and argue vigorously for retention even if it meant some sacrifices in other budgetary areas. In some instances spe- cific steps were taken to mobilize faculty support for such representations. 9 Location and Service Promotion The issues of location and service promotion are linked in an interesting way. The linkage is connoted in a nega- tive manner by the proposition that it is easy enough to hide the capability and thus inhibit demand. Such a cir- cumstance was aptly demonstrated by the installation in the computer science department at the University of Man- chester. No imaginative U.S. librarian would have much difficulty in achieving the same outcome, if for some peculiar reason it was desired. As indicated above, the dynamic upon which demand appears to depend is word-of-mouth promotion among facul- ty and students. In effect, a nucleus of satisfied customers spreads the word. To expand somewhat, the British ex- perience with promotional devices, such as printed announcements, was dismal. These usually cogent and articulate mes- sages were either ignored or misinter- preted. A perhaps sociologically reveal- ing misinterpretation that occurred at several sites was the erroneous belief that the service was limited to persons having faculty status. Word-of-mouth dissemination has to have an initial trigger. This means that in some instances the librarian or other staff member has to take the initiative to identify individual prospective cli- ents (e.g., students at a crucial stage in their thesis projects) and persuade them to try the system. Word-of-mouth also has its own time- lag pattern: the demand curve ultimate- ly can become exponential, but the ini- tial period always seems sluggish. The need to be able to exert some control over demand growth is illustrat- ed by the developments at Cranfield In- stitute of Technology. At the beginning of the academic year, demand was mod- erate, but enough clients were recruited to provide the trigger. Because of a rel- atively invariant activity sequence in their M .Sc. programs, the crucial stage of thesis planning occurred for a batch of 300-plus students all at the same time, just after the tum of the year. This period happened to coincide with the natural acceleration of the demand- growth curve. The service was limited to a capacity of two to three searches a day because of technicalities associated with the time differential between the U.K. and the U.S. Consequently, the service personnel experienced an acute overload situation during January and February and were forced to institute an advance-booking procedure. At one paint, reservations were being made three weeks in advance. Moreover, be- cause of activity deadlines, even the ad- vance-booking procedure left some prospective users out. Although it was not documented as such, one suspects that the circumstances produced some attitudinal backlash. It should be emphasized that there is no imputation here that the service was oversold. It was just that a combination of factors (including the relatively co- hesive characteristic of the student body) generated a singular peak in the demand pattern. The message for library management is clear: while a high level of demand, Bibliographical Services I 289 in effect, certifies the value of the ser- vice, some analytical effort is essential. Thus one can anticipate the pattern of demand growth and the rhythm of de- mand oscillations. Access procedures can be instituted to dampen the fluctuations and prevent any unfortunate whiplash effect from alternations between acute underuse and acute overload. To return to the locational issue in the physical sense, the main admonition relates to convenience for both the users and the library staff. Ideally, the instal- lation would have some perceptual prominence in the sense that library patrons could become aware of the re- source in the natural course of other transactions. Such an arrangement would support the trigger process de- scribed above. On the contrary, the in- stallation should be isolated from noise and movement distractions because con- siderable concentration is required for the effective use of the system. A good compromise might be a glass-walled space in the science reference area. Such a prescription anticipates the is- sue of organizational location. Some af- filiation with reference operations seems logical in light of the bibliographic nature of the service. However, there is at least one other important factor for administrators to consider, and that is document delivery. It is a characteris- tic of on-line systems to generate rather large bibliographies (e.g., 200-300 titles) from what appear to be relatively nar- row searches. The user can be admon- ished to be selective in his or her requests for full-text materials but even so will often put the resources of even the best collections to a severe test. The prescriptive inference from all this again seems clear: there must be an arrangement of close cooperation be- tween the managers~ of the on-line ser- vice and the staff responsible for delivery to the user of full-text items. In most instances, such arrangements should include interlibrary loan and referral functions. 290 I College & Research Libraries • July 1977 The penalties for inadequate plan- ning in this particular matter can be severe. Users can experience profound frustration if they are presented with their heart's desire in the form of an impressive list of patently relevant titles only to be told that some are unobtain- able or obtainable only after a delay. Another implication relates . to the British experience regarding those in- stallations not located in libraries. In such cases, the tenuous link between document identification and document delivery made the whole process of eval- uation somewhat ambiguous. That is, users could express a high degree of satisfaction with the bibliographic search process and the product bibliog- raphy, but there was no easy way of knowing whether the ultimate outcome in terms of information acquisition was even moderately satisfactory. Those di- rectly involved on the search side specu- lated openly that it probably often was not. Procedures of Use Finally, let us return briefly to the procedures-of-use issue. As suggested above, there are logical reasons to avoid the extremes of user participation in the on-line search process. This propo- sition implies that there might be an optimum middle ground. Indeed, the observer's personal preference is for a mode of use whereby the intermediary (librarian) operates the terminal and the user is present and can participate in the formulation and revision of the query and can observe the output. How- ever, in all fairness the ideal arrange- ment is more probably one which is flexible. That is, the service should probably be able to accommodate the user who is sufficiently expert to operate the terminal in a completely do-it-your- self mode as well as the user who would prefer to delegate the search completely to the information professional. Such a flexible procedure will require an adaptable and accommodating staff and a willingness to take the time to ne- gotiate with each user as an individual. Though such individualized service is relatively ·costly, it is certainly within the accepted ideology of the profession. REFERENCES 1. See, for example, R. B. Downs, "University of Illinois Library. Annual Report, 1970- 71," p. 7. 2. R. E. Hock, "Providing Access to Externally Available Bibliographic Data Bases in an Academic Library," College & Research Li- braries 36:208-15 (May 1975). 3. J. J . Gardner, D. Wax, and R. D. Morrison, Jr., "The Delivery of Computer-Based Bib- liographic Search Services by Academic and Research Libraries," ARL Management Sur- vey 2:1-6 (Sept. 1974). 4. Lockheed has announced the imminent availability of two new files, Historical Ab- stracts and America: History and Life, which would appear to portend significant capabili- ty in the general humanities area. 5. This figure is based on an interpolated esti- mate from several statistical sources, includ- ing the listings of institutions designated class I and II-A made by the American As- sociation of University Professors ( AAUP Bulletin 61:142 (Summer 1975)) and U.S. Office of Education statistics ( S. V. Smith, Library Statistics of Colleges and Universi- ties: Fall 1971: Analytic Report (Part C) (Washington: Govt. Print. Off., 1973). 6. Extrapolated from data presented in D. E. Williams, "Report to the Vice Chancellor for Academic Affairs on On-Line Literature Search Services" (College Park: McKeldin Library, University of Maryland, 1975). 7. R. K. Summit, personal communication, April1976. 8. H. Perkin, "The Financial Crisis in British Universities: Or How to Live with 29 Per- cent Inflation," AAUP Bulletin 61:304-8 (Winter 1975) . 9. J. S. Kidd, "Toward Cost-Effective Proce- dures in On-Line Bibliographic Searches," College & Research Libraries 38:153-59 (March 1977).