fosmire.p65 500 College & Research Libraries November 2000 Free Scholarly Electronic Journals: What Access Do College and University Libraries Provide? Michael Fosmire and Elizabeth Young Results of a survey of the level of access provided to selected, free scholarly electronic journals by academic libraries is reported. A list of scholarly journals that are available electronically without any access limitations was compiled, and for each title the number of holding institu­ tions on OCLC and its coverage by major abstracting and indexing ser­ vices were determined. The twenty-five most popular titles, determined by number of holding institutions on OCLC, were searched on a random sample of library Web sites and catalogs to provide a separate gauge for access levels provided by institutions. Despite approximately half the titles being indexed by major vendors, including 88 percent of the top twenty-five, access levels were fairly low and decreased dramatically with size of the institution. The Bulletin of the American Mathematical Society and Journal of Extension provide examples of free e-journals with print counterparts. Approximately one-half and one-third of the in­ stitutions, respectively, provide access to only the print versions of these journals (in the latter case, the print version is no longer produced, so access to current issues is not provided). his investigation was sparked by a presentation at the 1999 Special Libraries Association Annual Conference, where a representative from the Institute of Phys­ ics stated that “librarians seem to have trouble figuring out what to do with free things.”1 The statement was in reference to the New Journal of Physics, an electronic- only journal for which there are no sub­ scription fee and no access limitations. The society was wondering why libraries were so slow in cataloging the journal and mak­ ing it available to their users. Because the New Journal of Physics is a general physics journal published by a reputable source, it would appear to be in the scope of most libraries’ collections. However, a quick search on OCLC yielded only a handful of institutions that admitted to holdings of the journal. Although it could be argued that the journal had published only two articles at the time of the conference, the question raised is valid: To what extent are libraries providing appropriate access to free scholarly e-journals? Background There is no doubt that e-journals are here to stay. They offer more convenient navi- Michael Fosmire is the Physics and Earth and Atmospheric Sciences Librarian at Purdue University; e- mail: fosmire@purdue.edu. Elizabeth Young is the Assistant Coordinator, Technical Services, at SUNY Oswego: e-mail: eyoung@oswego.edu. 500 mailto:eyoung@oswego.edu mailto:fosmire@purdue.edu Free Scholarly Electronic Journals 501 gation between articles, faster time to publishing for authors, and enhanced multimedia content, to name a few ad­ vantages. As publishers more fully actu­ alize the e-journal, it soon will be as in­ sufficient to offer only print journals as it is to provide only print abstracts and in­ dexes. Hillary Tomney and Paul F. Bur­ ton found that their users had a gener­ ally positive attitude toward e-journals, with ease of access (especially from their desktops) being the format’s most impor­ tant perceived advantage.2 In addition to the above-listed advantages, 20 to 25 per­ cent also recognized the potential of hav­ ing attached comments and a broader dis­ semination of their work. This study provides one piece of the puzzle of “appropriate” access to free scholarly e-journals— amount of access. The growth in electronic journals has been dizzying. In 1994, Hans Roes was able to identify thirty-nine peer-reviewed scholarly e-journals.3 By 1996, Stephen P. Harter and Hak Joon Kim4 and Steve Hitchcock, Leslie Carr, and Wendy Hall5 had found 77 and 115 such titles, respec­ tively. In 1997, ARL’s Directory of Electronic Journals, Newsletters, and Academic Discus­ sion Lists (7th edition) claimed 1,002 peer- reviewed e-journals.6 This growth testi­ fies to the entrenchment of the medium in scholarly publishing. The overall electronic publishing en­ vironment appears to be healthy, with most major publishers now offering elec­ tronic versions of their journals, but what about the free electr onic journals? Hitchcock, Carr, and Hall found that forty-seven out of eighty-three STM (sci­ ence, technology, and medicine) titles in­ vestigated were free and planned to stay freely accessible.7 Looking at all disci­ plines, Harter and Kim found that almost 90 percent of e-journals were free.8 With the increase of commercial publishers in the e-journal market, the numbers have changed considerably, and the current investigation (based on the ARL directory listings) shows that only 213 (18%) out of 1,209 scholarly e-journals are free. However, the mere existence of free scholarly e-journals does not automati­ cally mean that libraries should provide access to them. On some level, the jour­ nals need to be of good quality and use­ ful. In general, libraries state in their col­ lection development policies that e-jour­ nals undergo the same selection process as print journals.9 Typically, selection cri­ teria are based on several factors, includ­ ing:10 1. importance of the periodical, as in­ dicated by its inclusion in standard bibli­ ographies and/or “recommended lists”; 2. importance of periodical’s subject matter in relation to the university’s edu­ cation goals and the curriculum; 3. inclusion of the periodical, and therefore access to its articles, in indexes and abstracts (both print and electronic) maintained by the library; 4. authoritativeness, as reflected by recommendation or adverse criticism by respected authorities; 5. reputation and standing of the pub­ lisher; 6. price. To determine the importance of e-jour­ nals, which provides some insight into the fourth factor above, Harter did an impact study.11 He found that of the thirty-eight e-journals that started publication before 1994, twenty-three were cited by other journals indexed in ISI Citation indexes, including seven that were cited more than eleven times. In addition, he found that the Bulletin of the American Mathematical Society was wildly successful, being cited more than 1,500 times in 1994. Public Ac­ cess Computer Systems Review (PACS-R) ranked in the 71st percentile in impact factor among journals in its subject area, and Psycoloquy achieved a 76th percentile ranking. Thus, some free e-journals have provided high-quality content to their disciplines. Because neither PACS-R nor Psycoloquy were themselves indexed by ISI, their showing was indeed impressive. However, Harter did note that the total number of articles written was rather http:study.11 502 College & Research Libraries small, so the overall contribution of the journals to the field was less. An impact study of the STM journals from this study’s sample found a sizable number of high-quality journals with higher total numbers of articles published.12 The third factor listed above was inves­ tigated in the present study, which found that 42 percent of the identified free schol­ arly e-journals were indexed by a major A+I service (see discussion section). This finding also lends credibility to these publi­ cations as a nontrivial part of the schol­ arly literature. Of course, the sixth factor above is not an issue in the present situa­ tion. The reputation and standing of the publisher (factor five) varies considerably by title, with the American Mathematical Society (Bulletin of the American Mathemati­ cal Society, Electronic Research Announce­ ments, Notices), Institute of Physics (New Journal of Physics), and the U.S. government (Emerging Infectious Diseases) being ex­ amples of highly reputable publishers pro­ viding free e-journals. Many of the free e- journals are hosted by professional societ­ ies, although certainly many others are hosted simply by a group of interested people working in the same subject area. The other criteria are beyond the scope of the present study, although with the wide variety of subject areas covered by the titles, most libraries would seem to have an interest in some titles (factor two). Harter ’s study and the encouraging level of indexing of free scholarly e-jour­ nals, along with this study’s authors’ ob­ servations on the other selection factors, provide solid motivation for looking at the level of access libraries provide for their users. This study provides one piece of the puzzle of “appropriate” access to free scholarly e-journals— amount of ac­ cess. Other research needs to be done to ascertain more fully the quality of the e- journals and thus whether the access re­ ported here is indeed appropriate. Methods Before proceeding with methods used in the study, the term free scholarly electronic journal must be defined because there are November 2000 certainly ambiguities in each word. For the purposes of this article, free refers to something that can be accessed by any­ one with an Internet connection, regard­ less of a subscription or membership sta­ tus with any professional society or publisher. Thus, all of the “free, with print” journals do not qualify. The word scholarly refers to the narrow concept of scholarly and peer-reviewed. For ex­ ample, the very popular and scholarly Issues in Science and Technology Librarianship—although very important to sci-tech librarians, indexed by A&I ser­ vices, and free—does not claim to be peer­ reviewed.13 Thus, it was not included in this survey. The criteria for a publication being scholarly, then, are that (1) articles have references and a scholarly treatment and (2) the journal identifies itself as be­ ing peer-reviewed. Electronic journal refers not only to electronic-only journals, or even electronic-first journals, but also to any journal that has an electronic version containing at least as much content as the print version and published approxi­ mately simultaneously with the print ver­ sion, if not earlier. For the purposes of libraries providing access to journals with both print and elec­ tronic versions, access is defined as access to the electronic version. Furthermore, the journal must be published currently, so only journals that have published a 1999 issue or have shown other signs of being active were included in this sample. This led, for example, to the exclusion of the Katherine Sharp Review and PACS-R, one of which announced it had ceased publica­ tion and the other of which showed no recent activity at the time of this study. Candidate journals were taken from the Directory of Electronic Journals, Newsletters, and Academic Discussion Lists, 7th edition. A search of peer-reviewed journals in all subject areas yielded 1,209 titles. Inspec­ tion of titles in the database to identify free e-journals narrowed the list to 373. At­ tempts were made to access the remain­ ing e-journals, given the Uniform Resource Locators (URLs) in the directory. The au­ thors assessed whether the journal was (1) http:reviewed.13 http:published.12 available from that URL, (2) free, (3) schol­ arly, and (4) still published. Some of the directory’s URLs led to sites with multiple titles, notably the EMIS site, which con­ tains more than twenty e-journals.14 The final number of free scholarly e-journals was 213, much larger than the forty-seven and sixty-seven quoted in the earlier stud­ ies. Table 1 contains a breakdown of titles by subject area. As with any undertaking this large, it is difficult to produce a definitive list of all the free scholarly e-journals. Correct identification depends on the directory having accurate information about the na­ ture of the journal (including its current link) and the journal’s Web site having accurate information. Although the au­ thors of this study cannot profess to have compiled a complete list of free scholarly e-journals, they did make every effort to compile as comprehensive a list as pos­ sible. This pool of titles includes, at the very least, a large majority of available titles and offers a good indication of the overall nature of the information land­ scape for this genre of materials. To gauge the overall level of access aca­ demic libraries offer to these titles, two methods were used. First, for all the quali­ fying titles, OCLC was searched to find the number of holding institutions (see table 2, column 1). However, it certainly is pos­ sible that institutions provide access to a title without registering it with OCLC; thus, a simple search on OCLC may un­ derestimate the total level of access actu­ ally provided. Possible reasons for not reg­ istering include the cost of formally cata­ loging a title and responsibility for filling interlibrary loan (ILL) requests. Thus, for example, a journal may be located on an “e-journals” page on the library’s Web site but not appear in the catalog. The second gauge was to look at a sample of institutions’ library Web sites and catalogs to determine whether the in­ stitution provides access to the title. Be­ cause preliminary searching indicated that the overall level of access for free e- journals was low, to maximize the chances of success in finding access, only the top Free Scholarly Electronic Journals 503 TABLE 1 Breakdown of Free Scholarly E-journals by Subject Area Subject Area % of Sample*(N) Physical sciences 39 (71) Math 28 (52) Arts and humanities 22 (40) Medicine 21 (38) Engineering/computers 12 (22) Education 10 (18) Social sciences 7 (13) Biological sciences 7 (12) Business 5 (10) Library science 3 (5) * The percentages add up to more than 100. because interdisciplinary titles were entered under multiple categories. twenty-five journals in terms of number of holding institutions in OCLC were searched. The institutions surveyed were se­ lected from the list in the Carnegie Foundation’s Classification of Institutions of Higher Education.15 The universities were selected at random with replace­ ment, so some of the institutions are weighted multiple times in the final analysis. First, the institutions were sepa­ rated by Carnegie classification type (combining institutions of different lev­ els [I+II]) so that samples of different- sized institutions could be analyzed. Then, a sample size (N) of fifty was cho­ sen for each classification level. For each institution selected, the library’s Web site was located and searches were done in both the online catalog and the rest of the Web site to lo­ cate access to the top twenty-five e-jour­ nals. If a title was listed in either the cata­ log or somewhere else on the Web site (typically, an e-journal page), it was con­ sidered accessible. Because only the most popular journals were searched, the re­ sults of this survey are an overestimation of access provided for free scholarly e- journals as a whole. Title searches for the http:Education.15 http:e-journals.14 TABLE 2 Top Twenty-five Free Scholarly E-journals by Number of Libraries with Holdings on OCLC Title Number % Research 1+11 % Doctoral 1+11 %Masters 1+11 1nde�ed 1n� of Holding 1nstitutions 1nstitutions 1nstitutions 1nstitutions Offering Offering Offering on OCLC Access Access Access Blood, Cells, Molecules, and Diseases 78 52 32 2 Medline, �oS, Chem A�s New Journal of Physics 69 66 24 2 INSPEC Educause Review 69 14 16 0 Electronic Research Announcements of the American Mathematical Society 56 50 14 4 MathSciNet Emerging Infectious Diseases 54 56 34 24 Medline, �oS, �SI Bulletin of the American Mathematical Society 41 64 44 12 MathSciNet Psycoloquy 38 34 10 4 Electronic Green Journal 32 28 20 6 PAIS, GSI, Env� A�s� Architronic 28 20 8 0 Avery Early Modern Literary Studies 28 16 16 6 MLA Education Policy Analysis Archives 28 36 12 2 ERIC Journal of Seventeenth-Century Music 28 32 14 2 � Journal de Theorie des Nombres de Bordeaux 28 14 0 0 MathSciNet Music Theory Online 27 20 16 4 RILM Conservation Ecology 25 24 14 6 Bio Abs� �oS� PAIS Optics Express 25 32 4 0 INSPEC� Chem Abs� �oS Geometry and Topology 24 30 10 0 MathSciNet Journal of Artificial Intelligence Research (JAIR) 24 16 20 0 MathSciNet� �oS Journal of Extension 24 34 14 0 Agricola� �oS� ERIC Journal of High-Energy Physics 23 28 6 0 MathSciNet� �oS Sociological Research Online 23 18 16 4 PAIS 504 C ollege & R esearch L ib raries N ovem b er 2000 � . Uo " oU U U� U 0 0 0 Free Scholarly Electronic Journals 505 journals on the institutions’ Web sites and in catalogs were done in January and February of 2000. Discussion Overall, the number of e-journals that showed holdings in OCLC was rather large. Of the 213 journals investigated, 56 (26%) had no holdings in OCLC , 70 (33%) were held by one to five institu­ tions, 36 (17%) by six to ten institutions, and 51 (24%) by more than eleven in­ stitutions. Holdings information for the top twenty-five e-journals is given in table 2. Another indicator of the rel­ evance of titles is their inclusion in ma­ jor indexing and abstracting services. Ninety-five journals were found to be indexed in a major A+I service, with MathSciNet indexing the most e-jour­ nals at forty-three (83% of the free math­ ematics e-journals in the study sample are indexed in this service).16 Of the top twenty-five journals by OCLC hold­ ings, 88 percent were indexed by a ma­ jor A+I service. The free scholarly e-journals repre­ sent all subject areas (listed in table 1). The sciences, engineering, and medi­ cine comprise more than 60 percent of the titles identified, but the arts and humanities, education, and the social sciences also are well represented. Simi­ larly, the top twenty-five journals cover all the major subject areas, from the STM literature to arts and humanities and the social sciences. Only business is not represented in the top twenty-five list. The authors’ in-depth analysis of the top twenty-five free scholarly e-journals showed that access to journals varies greatly by size of institution (see table 3). Research institutions provide access to roughly twice as many titles as doc­ toral institutions and nine times as many as master ’s-level institutions. Although some of this difference may be attributable to differences in the scope of the collections of the institu­ tions due to the broad distribution of subjects covered by the analyzed jour­ http:service).16 506 College & Research Libraries November 2000 TABLE 3 Average Number, Minimum, and Maximum of the Twenty-Five Free Scholarly E-journals to Which a Library Provided Access scriptions provide access to the online version, de­ spite there being no re­ strictions on access. In addition, no institution without a print subscrip­ tion provided access to the online version of the Bulletin. When such a mainstream free e-journal has such a hard time reaching the library com- Type of Institution Average Minimum* Maximum Research I+II 7.7 0 (4%) 23 Doctoral I+II 4.1 0 (28%) 18 Master's I+II .82 0 (56%) 6 *The percentages in the minimum column refer to institutions that provided access to none of the top twenty-five scholarly e-journals. nals, it is likely that at least some jour­ nals would support the curriculum or re­ search interests of the institutions. Nev­ ertheless, seven times as many doctoral institutions do not provide access to any of the top twenty-five journals, and four­ teen times as many master ’s-level insti­ tutions do not provide access to any of those titles as research institutions do. The Bulletin of the American Mathemati­ cal Society provides a striking example of the lack of institutional access to free scholarly e-journals. If an institution has one subscription to a mathematics jour­ nal, it is this one. And if the institution has a subscription to the print version, it follows that it also is interested in the online version (which is available free even without a print subscription). Yet, fully 50 percent of the master’s-level in­ stitutions have subscriptions to the print version, but only 12 percent also provide access to the electronic version (see table 4). Electronic access increases dramati­ cally with institution size, in the ratio of 1:3:1:0:0:3 for research, doctoral, and master’s-level institutions, respectively. Furthermore, the high level of elec­ tronic access to the Bulletin is helped by the URL to the online version in the OCLC record for the print product. Sev­ eral libraries provided access to the online version almost by default by having links in the print record (not all the URLs were linked, even though most of those cata­ logs had that capability). With that boost to the access rate, it is surprising that only 54 percent of institutions with print sub­ munity, the low rate of access to the electronic- only journals is easily understood. Some of the other journals also had access levels augmented by other than conscious selection decisions. Blood, Cells, Molecules, and Diseases is a title available on IDEAL through Academic Press (and several institutions volunteered that their access was through the IDEAL site and not the free site).17 The Institute of Phys­ ics is able to include the New Journal of Physics in its site-licensing agreement for its pay periodicals (even then, some in­ stitutions that had signed an agreement with IoP for their subscription journals did not provide access to the New Journal of Physics). Emerging Infectious Diseases is a government document put out by the Centers for Disease Control, and links to the online version were often off the gov­ ernment document print record. Thus, of the five journals with the greatest access TABLE 4 Print versus Electronic Access to the Bulletin of the American Mathematical Society Type of Electronic Print Institution Access (%) Access (%) Research I+II 64 92 Doctoral I+II 44 82 Master's I+II 12 50 Note: Numbers are percentages of institutions that provide access to the electronic version of the publication and those that provide access to the print version. http:site).17 Free Scholarly Electronic Journals 507 by the institutions sampled, only one, Electronic Research Announcements of the American Mathematical Society, was not artificially high either by bundling with other journals in a package or because of links from its print version. The Journal of Extension is an example of a print-based title moving to an en­ tirely online publication. In the case of research and doctoral institutions, 50 percent more (nine and three, respec­ tively) could have provided access to the online version if they had just created a link from the print record, which often contained the note: “Continued by online version.” Indeed, although no master ’s­ level institutions provided access to the journal’s online version, six percent pro­ vided access to the print version. The journal obviously had met the selection criteria for inclusion in the libraries’ col­ lections, but when it ceased to be issued in print, the libraries failed to offer ac­ cess to the electronic version in either their catalogs or Web sites. Not including the top five, access to the other top free e-journals, most of which are not well connected to professional society or commercial publishers, hovers around 20 percent for research institu­ tions, 12 percent for doctoral institutions, and 2 percent for master ’s institutions. Conclusion The genre of free scholarly electronic jour­ nals embraces all subject areas and ap­ pears to be growing steadily, although the level of access that libraries provide is lower than expected. Although almost half of the e-journals in this study’s over­ all sample and 88 percent of the top twenty-five titles are indexed by a major service, almost no master’s-level institu­ tions provide access to them. Moreover, even the typical research institution pro­ vides access to only about one-third of the top twenty-five titles. The rest of the free scholarly e-journals are even less acces­ sible through academic libraries. In this case, it appears that the indexing services are substantially ahead of the library com­ munity in recognizing the importance of these titles in the scholarly information landscape. The Bulletin of the American Mathematical Society and the Journal of Ex­ tension are prime examples of journals that libraries are not providing access to at a level commensurate with their appar­ ent value. Of course, the present study could not investigate all the factors involved in the selection of new periodical titles by aca­ demic libraries, but it certainly appears as though several worthy journals are not getting the attention they deserve from the library community. Further study, perhaps a survey of libraries asking how certain titles did not meet selection re­ quirements, may shed light on the low access rates found in this study. One might investigate whether, for example, the free scholarly e-journals are not meet­ ing the selection criteria, the stated fac­ tors for selection are not being used, or a different set of selection factors is being applied to such journals. Indeed, are there still technological barriers to providing access to free scholarly e-journals, espe­ cially among the smaller institutions? If libraries are to prevail in the serials crisis that has been prevalent for the past twenty years, librarians need to reward the journals that are trying to provide content at a reasonable price (in this case, free) by at least making them available in library catalogs and/or on library Web sites. The New Journal of Physics notwith­ standing, where the Institute of Physics aggressively marketed its free e-journal in order to get libraries to provide access, the authors would have to agree with the institute’s representative that, generally, librarians just do not know what to do with free things. Notes 1. Jerry Cowhig, comments at the 1999 Physics, Astronomy, Mathematics Vendor Update, 90th Special Libraries Association Annual Conference, Minneapolis. 508 College & Research Libraries November 2000 2. Hilary Tomney and Paul F. Burton, “Electronic Journals: A Study of Usage and Attitudes among Academics,” Journal of Information Science 24.6 (1998): 419–29. 3. Hans Roes, “Electronic Journals: A Survey of the Literature and the Net,” Journal of Infor­ mation Networking 2.3 (1994): 169–86. Also available online at: http://cwis.kub.nl/~dbi/users/ roes/articles/ej_join.htm. 4. Stephen P. Harter and Hak Joon Kim, “Electronic Journals and Scholarly Communication: A Citation and Reference Study,” Proceedings of the Midyear Meeting of the American Society for Information Science, San Diego, May 20–22, 1996, 299–315. Also available online at: http:// php.indiana.edu/~harter/harter-asis96midyear.html. 5. Steve Hitchcock, Leslie Carr, and Wendy Hall, A Survey of STM Online Journals 1990–95: The Calm before the Storm. Available online at: http://journals.ecs.soton.ac.uk/survey/survey.html. 6. Mogge, Dru W., ed., Directory of Electronic Journals, Newsletters and Academic Discussion Lists, Foreward. Available online at: http://db.arl.org/foreword.html. 7. Hitchcock, Carr, and Hall, A Survey of STM Online Journals 1990–95. 8. Harter and Kim, “Accessing Electronic Journals and Other E-Publications: An Empirical Study,” College & Research Libraries 57 (Sept. 1996): 440–56. 9. Elizabeth Parang and Laverna Saunders, Electronic Journals in ARL Libraries: Issues and Trends, SPEC Kit #202 (Washington, D.C.: ARL, 1994). The SPEC Kit includes copies of collection development policies for e-journals from seven major academic institutions. All of those policy statements claimed that selection policies were the same for electronic journals as they were for print journals. 10. Elizabeth Futas, Collection Development Policies and Procedures, 3rd ed. (Phoenix, Ariz.: Oryx, 1995), 306. The quoted passage is the collection development policy for serials at Hawaii Pacific University. All of the institutions in this text had similar selection criteria, with this example being the most succinct. 11. Stephen P. Harter, “Scholarly Communication and Electronic Journals: An Impact Study,” Journal of the American Society for Information Science 49, 6 (1998): 507–16. 12. Michael Fosmire and Song Yu, “Free Scholarly Electronic Journals: How Good Are They?,” Issues in Science and Technology Librarianship 27 (summer 2000). 13. At the time of this study, ISTL was not peer-reviewed. However, as of the summer 2000 issue, it will have a peer-reviewed section. 14. European Mathematical Information Service, “The Electronic Library of Mathematics: Mathematical Journals.” Available online at: http://www.emis.de/journals/index.html. 15. Carnegie Foundation for the Advancement of Teaching, A Classification of Institutions of Higher Education, 1994 Edition. Available online at: http://www.carnegiefoundation.org/Classi­ fication/index.htm. 16. To determine which A+I services index these journals, Ulrich’s International Periodicals Directory, 38th ed. (New York: Bowker, 2000) was consulted, as were the following indexes: MathSciNet, INSPEC, MLA, SocioFile, PsycInfo, CINAHL, Agricola, Medline, EconLit, Wilson’s: Humanities Abstracts, Social Science Abstracts, Business Abstracts, IIMP, PCI, GeoRef, ATLA, Business and Industry Index, Criminal Justice Abstracts, and Web of Science. Not all journals were listed in Ulrich’s, nor did Ulrich’s necessarily contain all the indexers of a journal. 17. Subsequent to the original submission of this paper, Blood, Cells, Molecules, and Diseases cut off free access to its most recent issues. Full text for the current five years is available through IDEAL only. http://www.carnegiefoundation.org/Classi http://www.emis.de/journals/index.html http://db.arl.org/foreword.html http://journals.ecs.soton.ac.uk/survey/survey.html http://cwis.kub.nl/~dbi/users