College and Research Libraries User Acceptance of Electronic Journals: Interviews with Chemists at Cornell University Linda Stewart Full-text electronic journals must be designed effectively if they are to support and promote scholarly activities. This paper presents interview data from chemists who have used an experimental electronic journal system. The first section explores the potential of electronic journals for accomplishing traditional scholarly tasks; the second focuses on the characteristics of an ideal electronic system and its effects on user pro- ductivity. Implications for system design are presented. Improvements must be made in the areas of portability, comfort, convenient access, permanence, and serendipity if electronic journals are to gain wide ac- ceptance. he success of scholarly elec- tronic journals depends on the users' ability to assimilate li!!!IJJI!!II~ them into their work habits. The ergonomic implications of electronic text have been reviewed by Andrew Dillon. 1 The effects of electronic publish- ing on scholarly communication have been explored by F. W. Lancaster2 and Don Schauder.3 To describe the interac- tion between electronic information and the intellectual processes supported by the journal literature, Jan Olsen inter- viewed forty-six academics who had never previously used electronic jour- nals.4 This paper presents interview data related to the potential utility of electronic journals by chemists who have used an experimental system of full-text journals. This information will assist in the effec- tive design of electronic systems so that they will support and promote scholarly activities. Both libraries and . individuals may someday, for economic reasons, need to choose between printed and electronic journals. Moreover, a comparison with printed journals forms a common baseline that users can refer to in describ- ing the potential of electronic journals. In this light, the first part of this paper ex- plores the potential of electronic journals for accomplishing the scholarly tasks tra- ditionally achieved using printed jour- nals. The second part focuses on selected features of an ideal electronic journal sys- tem and its potential influence on user productivity. Methodology The data derive from open-ended, hour- long interviews with thirty-nine users of Linda Stewart is the Bibliographer at the Albert R. Mann Library, Cornell University; e-mail: lgsl@cornell.edu 339 340 College & Research Libraries the Chemistry Online Retrieval Experi- ment (CORE) at Cornell University. The CORE system, described in detail by Ri- chard Entlich,5' 6 Michael Lesk/ Larry Krumenaker, 8 and Stu Borman/ consisted of the full text of twenty American Chemi- cal Society (ACS) journals and the corre- sponding scanned images of the source journal pages. These were loaded locally at Cornell and made available free to us- ers, twenty-four hours per day, in their offices and certain public buildings. Two different user interfaces were available, one with exact page images and the other with machine-readable text. The re- sponses analyzed in this paper do not Interviewers were encouraged to make questions open-ended and to probe for more information when appropriate. deal with the particulars of the CORE system, although, of course, they were influenced by it. Instead, this research was an attempt to persuade users with this defined experience of electronic jour- nal systems to speculate more widely on the genre. All interviewees were members of the Cornell community who had used the · CORE system at least once and who con- sented to a one-hour personal interview during May and June 1995. Thirteen li- brary staff members conducted the inter- views. They scheduled appointments, took notes during interviews, and input the responses on a shared electronic tem- plate. The responses were later merged to form a database. This research project was designed to gather information for the design of elec- tronic journals rather than to test prede- termined hypotheses. Survey questions, which were pretested prior to the survey, attempted to elicit not only opinions but also the reasons for them. Interviewers were encouraged to make questions open-ended and to probe for more infor- July 1996 mation when appropriate. Analysis of nonfactual responses involved grouping together all responses to a particular ques- tion, looking for trends within the re- sponses, and coding the responses ac- cording to categories generated by the responses themselves. Characteristics of Participants The thirty-nine interviewees included twenty-one graduate students (53.8%), fourteen holders of doctorates (35.9%), and four undergraduate students (10.3%). Doctorates included seven faculty mem- bers, one visiting faculty member, three research associates, and three postdoc-: toral fellows. Thirty-four users were male. Fifty-nine percent were affiliated with the chemistry department, and 10.3 percent each with materials sciences I engineering and food science. The remainder were distributed among a variety of depart- ments. The average number of search ses- sions per user was 20.6, with a range of 1 to 222. Comparison of Electronic Journals with Printed Journals System Features Needed for Journal Use Olsen, from her interviews with scholars, identified a list of functions that scholars considered essential in an electronic sys- tem.10 These functions fell into two cat- egories: those that aid in selecting opti- mal articles and those that facilitate read- ing them. Activities for selecting the op- timal literature include: • browsing text to support ongoing education and generate new ideas; • browsing graphics to determine the value of an article; • flipping pages and scanning;. mak- ing physical contact with the material being read; • experiencing serendipity; • having a visual overview of a wide body of literature. Features facilitating reading include: • physical comfort; User Acceptance of Electronic Journals 341 • adequate type fonts, text design, and layout; • creation of a print copy; • portability of the text; • capability of underlining and an- notating. Interviewees were asked to evaluate the importance of each of the above fea- tures. Answers were coded as "very im- portant," "important," and "not impor- tant" by the author. Table 1 lists the tasks and features in approximate order of im- portance to the participants. All the features except two were con- sidered at least "important" by a major- ity of users. The single most important feature was the ability to create a printed copy; almost everyone considered it "very important." Four users mentioned the ability to annotate; four others, the abil- ity to preserve materials; two, portabil- ity; and two, physical comfort. Printing also allows the comparison of articles and is important for use at the benchtop. One respondent used printing to record the fact that an article had been read. The ability to browse gr_aphics and text also was of great importance. Users val- ued browsing graphics for speed and for the fact that a particular discipline may be better expressed graphically. Browsers seek particular structures that are more easily spotted than words within text. Macromolecules may be structurally com- plex and require color-coding. Browsing text was important for both generating new ideas and keeping cur- rent with related research. The impor- tance of browsing to save time also was suggested by ADONIS users in Great Brit- ain, as described by John Richardson. On 154 occasions (75.12 %), users looked only at the first page and then made a deci- sion as to printing the article. 11 Portability of the text, considered by most respondents to refer to portability of the physical copy as opposed to port- ability of the information, also was con- sidered key. This is understandable within thecontextofD. J. Pullinger's 1983 investigation, which showed that among potential users of the BLEND electronic text system in the United Kingdom, 65 TABLE I Importance of Tasks and Features to Scholars (by percentage of usable responses) Task or Feature Very Important Important Creation of a print copy 80.0% 14.3% Browsing graphics to determine 72.7 15.2 the value of an article Browsing text to support ongoing 65 .8 21.1 education and generate new ideas Portability of the text 52.6 31.6 Flipping pages and scanning 44.7 39.5 Physical comfort 36.8 44.7 Adequate type fonts , text design, 29.7 56.8 and layout Capability of underlining and annotating 40.5 32.4 Experiencing serendipity 44.4 25.0 Making physical contact with the material 13.5 18.9 Having a visual overview of a range 8.1 27.0 of literature Not Important 5.7% 12.7 15 .2 15.8 15.8 18.4 13.5 27.0 30.6 67.6 64.9 342 College & Research Libraries July 1996 TABLE2 Potential Satisfactory Performance by Electronic Journal System percent of bro:vsing and 85 per- cent of reading of entire papers took place outside office hours. 12 Annette Simpson found similar high percentages for twenty aca- demics in a variety of disci- plines .13 Visual overview of a wide body of literature and physical contact with the article were the two features considered · Task or Feature % ofYes or Maybe Responses Adequate type fonts, text design , and layout Browsing text 91.9% 90.6 85.3 ·least important by users. Al- though one interviewee felt that seeing a variety of journals might provoke interest, others men- tioned that they already knew which journals they wanted. Browsing graphics Capability of underlining and annotating 84.8 82.9 80.6 78.4 73.7 56.3 A few participants valued physical contact with the article: Flipping pages and scanning Portability of the text Physical comfort Experiencing serendipity Creation of a print copy An advantage of print is that I can use my finger to mark my place if I'm interrupted while reading .... With print, I feel that I'm more in "reading mode" whereas computer monitors are more like TV screens. I feel more passive; it's harder to read and pay attention. It makes you feel connected. Other- wise, the article goes away when you turn the machine off. It invokes the tradition and culture of the book, and I'm not comfort- able interacting with the screen. The feel and the color of the print are important. However, most negative responses em- phasized greater interest in the content of an article than in the reading process. Because most respondents considered physical contact with the article and wide visual overview "not important," these have not been included in the discussion below. Satisfaction of Requirements by Electronic Journals For each task or feature, users were asked if they felt that electronic journals, in CORE or any other system, could fill that need as adequately as print journals. (Table 2 lists the percentage of respondents an- swering yes or maybe for each task.) At least 84 percent or more of the re- spondents considered electronic journals capable of satisfying the need for ad- equate type fonts, browsing text and graphics, and underlining and annotat- ing. Slightly fewer respondents thought that electronic systems, with substantial modifications, could evolve so as to pro- vide adequate flipping and scanning abilities, and portability. Physical comfort and experiencing serendipity were more problematic, and almost half the respon- dents would still create a print copy of articles they considered important. Adequate type fonts, text design, and layout were generally considered within reach of electronic journals. The possibil- ity of personalizing the interface also was mentioned. Indeed, Peat Marwick ac- countants tested by Brewster Kahle et al. 14 wanted the ability to change the style or font themselves. Users felt that electronic journals could support the need for browsing text by providing speed, currency, and compre- hensiveness of coverage. User Acceptance of Electronic Journals 343 The notion of browsing graphics elec- tronically was attractive: Chemists are lucky; they have a sys- tem of drawing which is standard- ized. The minimum requirement would be to be able to draw a pic- ture, via something like Chem- Intosh, and search for it. The best would be to draw two structures and a transformation from one to the other and be able to search for · that. Users felt that electronic systems could provide annotating capabilities: to high- light text, cut and paste, insert notes, and even draw. Users provided a wealth of suggestions for facilitating flipping pages and scanning. It would be useful to have a map showing where you are in the ar- ticle-a square box with cells show- ing where you are .... You could click on the map to get where you want. If I could view two pages at a time, because we are used to physical page turning. I envision the following, which I would call a ''blank book": if they could make screens thin and the same size as pages, maybe out of liq- uid crystal, and manufactured a book of ten or so of these blank pages, you could then plug in a jour- nal article and then flip through. You could do something like "eye tracking," which is used in systems for paralysis victims. Laser-oriented ... tracks the eye's motions and per- forms certain tasks like turn pages when a certain area was looked at. However, others felt that flipping "pages" electronically is too slow. Many users felt that they could satisfy the need for portability of the text with electronic journals. A few (not included in the percentages in table 2), specified that they would make printed copies, which would then be portable. Others mentioned using a small workstation, being able to access the system from many locations, and saving articles to disk to be used in a variety of locations. Others were less satisfied: I want to be able to be reading while doing something else at the same time. For example, watching over a lab experiment. Laptops ... would be awkward to hold. Almost half the users found that us- ing the CORE system was as comfortable as using print: When you read in the library, you have to sit in one place, strain your eyes the same. The degree of comfort for CORE journal use was the same as for any other format. I would still have to get up to move around or get a glass of water. Another quarter of the users could imagine the potential for physical com- fort in reading electronic journals: It might help to be able to look down at a screen. That is a more natural reading position. [If it were] relatively the same size as paper. An interface that is more convenient than a mouse and key- board, like a light pen. However, a full quarter of users still felt that electronic journals would never be as comfortable to read as printed jour- 344 College & Research Libraries nals, mentioning eyestrain and lack of physical mobility while reading. The notion of experiencing serendip- ity while using an electronic system elic- ited diverse responses. Ten users men- tioned the search capability, a quality that distinguishes electronic systems from print, as an aid to serendipity: Even though you don't have the physical volumes next to each other the way you would in the stacks, you can get a collection of articles together that you can look at. If I type in "Kramer," I will get not only the articles written by Kramer but also the articles that cite Kramer. That can be an important associa- tion. Interviewers deliberately phrased the question to ask whether elec- tronic journals can ever replace the need for printed copies. Almost half the respondents answered nega- tively, ... In contrast, some respondents consid- ered the search capability a limitation be- cause it retrieves only what is requested. Other users mentioned different con- cerns: It seems to depend on how free you can let your imagination be, and print just seems better for that. There is a level of abstraction [with the computer] that interferes with my understanding of [the content]. Moving the mouse to highlight something breaks the concentration. In discussing creation of a print copy, because obviously electronic journals can be created that produce printed copy, the interviewers deliberately phrased the question to ask whether electronic jour- nals can ever replace the need for printed July 1996 copies. Almost half the respondents an- swered negatively, citing the lack of port- ability, the need for permanence or a physical record, the need to make trans- parencies or handouts, and the lack of physical comfort: No .... When I'm writing a paper, I browse through lots of things and like to have the papers arranged on my desk. On the screen . . . I can have only one or two papers open. No .... You can scan more articles faster [with electronic journals], but the only way to read them is to print them. These results resemble those of other studies, where large numbers of users prefer printed copies. Describing the ADONIS system, Richardson com- mented, "None of the users appeared to have read articles on screen instead of printing copies." 15 Schauder's survey of more than 700 academics from Australia, the United Kingdom, and the United States found that 75 percent would prefer to read ar- ticles as printouts (assuming laser-qual- ity printing) than on screen. 16 In general, comments from CORE par- ticipants agreed with interviews by Ann Bishop of eighteen engineering faculty members and students.17 The engineers needed the ability to search and display sections of articles, jump to cited links, customize interface features, retrieve and skim figures, download and manipulate (e.g., annotate) retrieved material, and experience serendipity. Substitution of Print Journals by Electronic Journals Participants were asked whether elec- tronic journals could be used for all schol- . arly journal reading. (Their answers are presented in table 3.) Negative responses were problematic. Although users were directed to consider all possible electronic User Acceptance of Electronic Journals 345 TABLE3 journal systems, and not just CORE, several cited CORE's insufficient updating or journal coverage, dem- Belief That Electronic Journals Could Be onstrating that they were not consid- Used for All Scholarly Journal Readin~ ering the entire range of systems. In Response table 3, the negative responses are % of Responses separated into two categories, de- Yes pending on whether the reasons Maybe 41.0% 12.8 given were specific to CORE rather No, for reasons than to electronic systems in g'eneral. specific to CORE 12.8 30.8 2.6 Nevertheless, more than half there- No spondents thought that electronic Answer unclear journals could possibly be used for all scholarly journal reading, whereas al- most one-third disagreed. Most positive responses stressed that such a system must be updated frequently and cover all the needed journals, and that the user must have convenient access to a termi- nal. The reasons for a negative response re- sembled those for needing a print copy: a need for permanence, portability, and ease of reading; and the instability of elec- tronic information. One professor cited: ... the risks involved in losing what's there. Systems go out of style and the longevity of electronic for- mats is dubious. Electronic informa- tion has a poor track record in this regard. To check consistency of responses, us- ers were asked how many ACS journals they subscribed to personally and whether they would drop these subscrip- tions if the journals were made available electronically on a permanent basis. How- ever, only sixteen respondents subscribed to any ACS journals, and only two sub- scribed to more than two. These low num- bers may be due to the student status of many of the interviewees. In Olsen's 1992 study, chemists subscribed to an average of five journals each, many of which were probably ACS journals.18 Of the sixteen subscribers here, four would discontinue their print subscrip- tions and eleven would not. (One person was unsure.) Three users who would not change commented that they disliked reading articles from a screen. Other com- ments were: It is more convenient to reach for the paper copy on my shelf than to open the system, wait, etc. I want a personal library of journals central to my work. Interestingly, three nonsubscribers vol- unteered that an electronic system might encourage them to subscribe to more jour- nals because of increased familiarity with them. The idea that print journal use might remain constant, or even increase, when the same journals are available electroni- cally was supported by preliminary data from the TULIP project, as described by Katherine Willis et al. 19 During the aca- demic quarter following the availability of various Elsevier /Pergamon journals online, use of the unbound paper issues of these journals rose more than 18 percent. Selected Characteristics of an Ideal System Users described their preferences in the areas of dates of coverage and presenta- tion of graphics and text. Dates of Coverage Three questions were posed concerning dates of coverage: the frequency of up- 346 College & Research Libraries dating necessary for readers to keep cur- rent, the number of past years needed for readers gathering background on a topic, and the number of past years available for locating specific articles or facts. Ques- tions were open-ended; standard dates or intervals were not suggested. Listed in table 4 for each question are the median, the range and the most common re- sponses. David Everett, in his study on using existing online journals to fill document- delivery requests at Stetson University, cited problems of insufficient periodical coverage and missing backfiles for many periodical titles. 20 In this study, most scholars desired backfiles covering thirty, or even fifty years, with updates every two weeks or less. However, some re- jected older information. Others claimed that the answer depended on the area of study. Presentation of Graphics Users were asked if they needed exact page images of graphics or whether ex- tracted graphics (available in one of the CORE interfaces) were adequate. The wording of this question inadvertently confused two issues: exact versus nonexact images, and proximity of the graphic to its context. However, because respondents were encouraged to explain July 1996 their answers, it was possible in many cases to determine which issue influenced the response. (See table 5.) The majority of users felt that extracted graphics were adequate or preferred; however, a significant number felt the need for context-text and graphics pre- sented in close proximity. This contrasts with Andrea Keyhani' s findings from the design phase of the GUIDON interface. 21 Her medical researchers preferred to view graphics in separate windows from the text. Presentation of Text Participants were asked whether they preferred exact page images or machine- readable text for textual material. Users preferring exact page images (35.1 %) slightly outnumbered those pre- ferring machine-readable text (29.7%) and those with no · preference (24.3%). Other preferences were for both options or were unclear (10.8%). Few reasons were cited for needing exact page images. Perhaps users believed, with Ann Schaffner, that: The structure of documents serves to orient readers; readers must be able to move from one section of article to another easily, and they must be able to locate and reread sections efficiently. 22 TABLE4 Desired Dates of Coverage for an Electronic System Question Median Frequency of updates Every 2 Weeks Backfiles needed for 20-30 years background research Backfiles needed for locating specific articles or facts Mid-1960s, or 30 years Range From 2 days before the print journal to once per year 2 years to the same as print Several years to the beginning of each journal's existence Most Common Response "As often as print," "Monthly" "The same as print" 40-50 years User Acceptance of Electronic Journals 347 Reasons cited for preferring machine- readable text included the ability to ma- nipulate text, search for words within text, and move text. Schauder, in his survey of more than 700 academics in Australia, the U. K., and the U. S., asked a similar question. 23 Re- taining the presentation style of the original article (de- More than half the respondents (52.6 %) said they would read more complete ar- ticles using an electronic system. The same number of articles or fewer would be read by 34.2 percent; other answers were unclear. Reasons for reading more articles included retrieving articles un- TABLES fined as same typography Graphics: Preference for Exact Page and layout) was considered Ima2es Versus Extracted Graphics "necessary" by 17 percent · I=====~=======================~====== and "desirable" by 56 per- Response % of Responses cent, compared to the prefer- f----'~--------------~--- ence for exact page images by only 35.1 percent in the present study. However, his question did not suggest the alternative of machine-read- able text. Extracted graphics are adequate. Text and graphics should be together, not necessarily exact images. 43.2% 16.2 27.0 Original page image is needed. No preference, mixed responses, unclear answers 13.5 Related to the need for exact page images is the desire for repro- duction of the familiar journal format. Participants were not asked to comment directly on this question, and outside opinions and studies vary. Schaffner24 hy- pothesized that: The grouping of articles into dis- crete journal titles with distinctive identities, and the format of indi- vidual articles, add structure and meaning to the body of scientific knowledge in subtle ways that are not yet thoroughly understood. However, in Schauder's study, a ques- tion on the importance of reading an ar- ticle as part of an entire journal issue, as opposed to an individual photocopy or reprint, elicited the response of "not im- portant" from 54 percent of the partici- pants. 25 Increased Productivity Users were asked whether they would read a greater number of complete articles using an electronic system than they read previously and whether they would read articles sooner. available in the library and devoting the time to reading that formerly was spent searching. By contrast, one respondent would read fewer complete articles, but would browse more. The majority of users (60.5 %) felt they would read articles earlier using an elec- tronic system, with 26.3 percent experi- encing no change and 13.2 percent being unsure or giving a mixed response. Posi- tive responses mentioned reasons of con- venience: I'm more of a night p er son a nd the libraries are not open that late. I like to be able to read at 2:30 a.m. Late at night, there are fewer dis- tractions-everybody's asleep, there are no classes, the restaurants are all closed, there are no sports going on. There are fifteen people sharing one journal in my lab .. . with a com- puter system, everyone has one. One negative response stressed: I generally read articles when I can find the time. 348 College & Research Libraries Conclusions Research on a rapidly changing field such as electronic information produces a snapshot of only one moment in time. In- terviews such as these provide a glimpse into the state of mind of potential users of electronic systems. The result is a prod- uct of the interaction between the state of the art of the products themselves, the current work habits of researchers, and their technological sophistication. The fol- lowing conclusions can be drawn from these interviews: • Of all the capabilities examined here, the abilities to create a printed copy and to browse both text and graphics were most important. • Users anticipated that most func- tions of a journal system could eventu- ally be accomplished by electronic jour- nals. However, a significant group of us- ers was pessimistic about their ability to experience serendipity, read comfortably, and forego printed copies. • About one-third of users did not anticipate ever replacing print journals entirely. • The number of personal print sub- scriptions would not immediately be af- fected by access to electronic journals. • For most users, systems should be updated at least biweekly and include back issues covering the past forty to fifty years. • Extracted graphics were adequate for most users, especially if they could also be viewed in context. • Users were slow to see the advan- tages of machine-readable text. • As a result of access to electronic journals, users expect to read more com- plete articles, spend their reading time more efficiently, and read articles sooner after publication. July 1996 As Schaffner commented: Electronic journals must, at the start, at least serve the basic functions that print journals have traditionally served. Once the transition has been made, new technologies may allow us to add new roles, to drop some of the traditional roles, or to fill them in intrinsically different ways. 26 Andrew Dillon's 1988 paper enumer- ates some of the possible enhancements: improved navigational capabilities, rapid browsing facilities, and customized print- ing Y Another enhancement, full-text searching, which was not addressed in detail in the CORE survey, already exists. The results of the present survey indicate that with greater experience using elec- tronic systems, chemists as a group are fairly optimistic about their potential. However, systems designers must still be concerned with improving the areas of portability, comfort, convenient access, permanence, and serendipity if electronic journals are to gain acceptance as a re- placement for print. Users still need to read comfortably from diverse locations, feel confident that online information will not "disappear," and make the chance connections with literature that spark creativity. In addition to meeting these exacting use requirements, electronic journal producers face significant chal- lenges outside this paper's scope in the areas of pricing, intellectual property rights, and conversion of a critical mass of information to electronic form. Despite the difficulties faced by database produc- ers and system designers, the CORE user study indicates a likelihood of user ac- ceptance if these challenges can be over- come. Notes 1. Andrew Dillon, Designing Usable Electronic Text: Ergonomic Aspects of Human Information Use (London: Taylor and Francis, 1991). User Acceptance of Electronic Journals 349 2. F. W. Lancaster, "Attitudes in Academia toward Feasibility and Desirability of Networked Scholarly Publishing," Library Trends 43 (spring 1995): 741-52. 3. Don Schauder, "Electronic Publishing of Professional Articles: Attitudes of Academics and Implications for the Scholarly Communication Industry," Journal of the American Society for Infor- mation Science 45 (1994): 73-100. 4. Jan Olsen, Electronic Journal Literature: Implications for Scholars (Westport, Conn.: Mecklermedia, 1994). 5. Richard Entlich, "Electronic Chemistry Journals: Elemental Concerns," Serials Librarian 25 (1995): 111-23. 6. Richard Entlich, Lorrin Garson, Michael Lesk, Lorraine Normore, Jan Olsen, and Stuart Weibel, ''Making a Digital Library: The Chemistry Online Retrieval Experiment," Communica- tions of the ACM 38 (Apr. 1995): 54. 7. Michael E. Lesk, "Electronic Chemical Journals," Analytical Chemistry 66 (July 15, 1994): 747A-55A. 8. Larry Krumenaker, "Virtual Libraries, Complete with Journals, Get Real," Science 260 (May 21, 1993): 1066--67. 9. Stu Borman, "Advances in Electronic Publishing Herald Changes for Scientists," C&EN (June 14, 1993): 10-23. 10. Olsen, Electronic Journal Literature, 29-53. 11. John Richardson, "Library User Trials with a CD-ROM Database," Computers in Libraries (Apr. 1991): 28-34. 12. D. J. Pullinger, "Attitudes to Traditional Journal Procedures," Electronic Publishing Review 3 (1983): 213-22. 13. Annette Simpson, "Academic Journal Usage," British Journal of Academic Librarianship 3 (spring 1988): 25-36. 14. Brewster Kahle, Harry Morris, Jonathan Goldman, Thomas Erickson, and John Curran, "Interfaces for Distributed Systems of Information Servicers," Journal of the American Society for Information Science 44 (1993): 453-67. 15. Richardson, "Library User Trials with a CD-ROM Database," 31. 16. Schauder, "Electronic Publishing of Professional Articles," 91. 17. Ann Peterson Bishop, "Scholarly Journals on the Net: A Reader's Assessment," Library Trends 43 (spring 1995): 545-70. 18. Olsen, Electronic Journal Literature, 16. 19. Katherine Willis, Ken Alexander, William A. Gosling, Gregory R. Peters Jr., Robert Schwarzwalder, and Beth Forrest Warner, "TULIP-The University Licensing Program: Experi- ences at the University of Michigan," Serials Review 20 (fall1994): 39-47. 20. David Everett, "Full-Text Online Databases and Document Delivery in an Academic Li- brary: Too Little, Too Late?" Online 17 (Mar. 1993): 22-25. 21. Andrea Keyhani, "The Online Journal of Current Clinical Trials: An Innovation in Elec- tronic Journal Publishing," Database 16 (Feb. 1993): '14-15,. 17-20, 22-23. 22. Ann C. Schaffner, "The Future of Scientific Journals: Lessons from the Past," Information Technology and Libraries (Dec. 1994): 239--47. 23. Schauder, "Electronic Publishing of Professional Articles," 91. 24. Schaffner, "The Future of Scientific Journals," 245. 25. Schauder, "Electronic Publishing of Professional Articles," 90. 26. Schaffner, "The Future of Scientific Journals," 240. 27. Andrew Dillon, John Richardson, and Cliff McKnight, "Towards the Development of a Full-Text, Searchable Database: Implications from a Study of Journal Usage," British Journal of Academic Librarianship 3 (spring 1988): 37-48. A world of information online THE FASTEST GROWING END-USER DATABASE SERVICE IN THE LIBRARY COMMUNITY • Full-text and images* Online • Web Access • NetFirst~ the authoritative guide to ·Internet resources • WorldCaf; the end-user version of the OCLC Online Union Catalog • More than 50 databases To t1nd out how FirstSearch® can help you move to the next stage of the electronic library: Call us at 1-800-848-5878, ext. 6425 to request a copy of the new FirstSearch brochure . Or view and order it online through our home page on the World Wide Web: http.//www.oclc. org *Coming soon .1111 fURTHERING AcCESS TO THE WoRLD ' S INFORMATION