Perceived Successes & Failures of Science & Technology E-Journal Access Previous   Contents   Next Issues in Science and Technology Librarianship Summer 2002 DOI:10.5062/F4MK69VR Perceived Successes and Failures of Science & Technology E-Journal Access: A Comparative Study Subject & Bibliographic Access to Science Materials Committee Science & Technology Section Association of College and Research Libraries Co-chairs: Janice Christopher Ibironke Lawal Steven J. Riel sriel@oeb.harvard.edu Members: Tom Auger, Sophie Bogdanski, Marian Burright, Bao-Chu Chang, Kavita Mundle, Anthony Oddo, Irwin Weintraub Abstract The ACRL/STS Subject & Bibliographic Access to Science Materials Committee undertook a study to identify perceived strengths and weaknesses of current access methods to peer-reviewed electronic journals in the fields of science and technology. This study focused on the perceptions of public-services librarians. Data were gathered by means of a questionnaire that the Committee posted on the World Wide Web. A request for participation was sent out through professional lists relating to science and technology librarianship, reference, and cataloging. The data provided by respondents indicated that most institutions provide more than one method to access their full-text e-journals. Public-services librarians were not totally satisfied with any of the five methods being surveyed, although the web-based listing method, which was the most used method, had a slight popularity over the others. Although over 70% of respondents' institutions used catalog records to provide access to e-journals, this was not a preferred method of access. No respondent preferred single linking to an aggregator's web site. Introduction In Spring 1999, the ACRL/STS Subject & Bibliographic Access to Science Materials Committee published a bibliographic essay identifying the main issues in the dissemination and use of peer-reviewed electronic journals in the fields of science and technology: access, pricing, cataloging & indexing, archiving, and licensing (ACRL/STS Subject & Bibliographic Access Committee 1999). Those topics are still very much issues of concern for librarians, library users, and publishers due to constantly changing commercial and technological environments. As the charge of the Committee is "to provide a forum for exchange of information relating to bibliographic access to science and technology materials," we chose to continue our previous work by focusing on the issue of access. We identified several different methods currently in use by American libraries for providing access to electronic journals. These methods are: 1) creating a bibliographic record which only describes the electronic version of a journal (i.e., a record that is separate from one for the print version of the journal); 2) using a single record to stand for both versions by adding URLs and access points for the electronic version to the record for the print version; 3) through web-based HTML lists; 4) through aggregator web sites; and 5) through database-driven (dynamically generated) HTML lists. For the purposes of this study, our assumptions about the current state of access to electronic journals are: 1) libraries have often initiated more than one access method in order to balance user needs with staffing, workflow, and local bibliographic control demands; 2) the responsibility for providing access may be distributed throughout the library system; and 3) the fast-paced changes in the current technological environment create confusion for both end users and librarians. Given the instability of this environment, the objective of this study is to identify the strengths and weaknesses that current access methods present to e-journal usability and maintenance. This study is not designed to address e-journal access issues from the user perspective, but rather to understand the benefits and limitations of each access method from the science and technology public-services librarian's perspective, based on his or her personal experience. Seeking to understand the experience of public-services librarians in assisting patrons with access to e-journals serves to begin a dialogue among parties responsible for access: public services, technical services, information technology, and possibly others. We wish to identify ways in which to minimize costly librarian effort while maximizing effective e-journal access. The questions under study are as follows: What are the most commonly used methods of access to peer- reviewed e-journals in the fields of science and technology at present? What do reference librarians at science and technology libraries perceive as the strengths and weaknesses of each of these methods? Are there less widely used methods? What are their strengths and weaknesses compared to those of the commonly used methods? Are there optimum methods of access in a given library context? By launching a qualitative study, we hoped to find some answers. Literature Review Beginnings When Woodward and McKnight (1995) wrote about issues of access and bibliographic control for electronic journals, many of the developments we now take for granted were just on the horizon. Bibliographic control of these resources largely amounted to directories, both print and electronic, alerting librarians and potential users of their existence. Both commercial and noncommercial publishers were relocating their e-journals to the World Wide Web. The MARC field of 856 had been proposed, and the Uniform Resource Locator (URL) had recently been approved. Virginia Polytechnic Institute and State University, Blacksburg was provisionally using the MARC fields 500, 590, and 866 in bibliographic records in their catalog to provide access to e-journals in the interim period before solutions involving the 856 field were approved and implemented. What kind of access is desirable? Hruska (1995) discussed two major obstacles to providing bibliographic control of electronic journals: the inadequacy of AACR2 for the description of such resources, and the inability of then-current OPACs to provide either relevant management-control systems or links from bibliographic records to related versions of a work. She noted that libraries traditionally viewed the catalog as an inventory of owned items and therefore segregated to specialized web pages the means of access to selected-but-not-owned Internet resources. Hruska warned that this approach was doomed because some print serials would cease publication and be succeeded by electronic serials; then librarians would want to link the closed-off record for the print serial in the OPAC to the URL for its online successor on their library's web page. Beverley Greer, as reported by Simser (1998), highlighted the importance of the library catalog, calling it the point where the library and the Internet intersect. She also called for the revision of AACR2 and MARC to accommodate the cataloging of electronic serials. Morgan (1995), Sleeman (1995), and Jones (1995) also argued for cataloging Internet-based electronic serials in the OPAC. Not everyone was convinced that this was the correct choice: Reynolds (1995) of the National Serials Data Program at the Library of Congress asked whether those who used the catalog to point to items not held by the library were confusing the function of a catalog with that of a bibliography, and whether there were not other ways to provide such links outside of the catalog, such as electronic directories with "hot links." Rich and Rabine (1999) examined the e-journal web pages of 114 academic libraries in 1998. After detailing the results of their survey, they offered four recommendations: 1) an e-journal web site should be linked from more than one page within the library's web site (and the names of the links should be clear, consistent, and descriptive); 2) search engines on a library's web site should be capable of finding individual e-journal titles listed there; 3) annotations for individual e-journal titles should be uniformly formatted and uniformly full whenever possible; and 4) librarians should create manageable web pages and be sure that the individuals responsible for them have the resources to maintain them effectively. Amanda Xu, as reported by Rodgers (1999), suggested that libraries add a trials page to their e-journals web page to highlight new resources being considered for purchase. In this way, libraries would offer their users access to a new category of materials: those under consideration for longer-term availability. Shirley Baker, as reported by Foster (1998), stated that, for the sake of remote users, librarians should make how to use their online systems (such as catalogs and web pages) so self-evident that training is not really necessary; in other words, not only must all of the information needed to access electronic serials be provided to the library user, but the meaning of that information must be made transparent. How is access actually being provided? Moothart (1996b) described a variety of early efforts to provide access to e-journals via library web sites at four U.S. universities. Three of the four provided added value to their alphabetical lists of e-journal titles by adding annotations about each title. One library did not provide a subject index to the e-journals listed on its web site; one library merely provided links to external web pages with lists of e-journals or articles in various disciplines rather than providing links to those e-journals to which it subscribed. Three years later, Rich and Rabine (1996) would recommend such an approach to libraries that do not have adequate resources to maintain their own lists of e-journals. A year later, Moothart (1997) outlined the decisions made at Colorado State University's Morgan Library regarding access to e-journals which had been donated after flooding had damaged many of its print serials. Faced with the enormous task of quickly providing access to numerous titles, the Library opted to create an alphabetical master list of e-journals on its web site while exploring the feasibility of a searchable list. Jennifer Edwards and Amanda Xu, as reported by Rodgers (1999) described how e-journals were handled at MIT around 1997. Formal coordination had been established between acquisitions and cataloging staffs, with the digital resources librarian in the acquisitions department responsible for maintaining an online list of full-text e-journals. Communication between the serials-cataloging unit and either public-services staff or users about e-journals, however, was still informal and incomprehensive: the serials-cataloging unit relied on informal notification from public-services staff and "on the kindness of strangers" about changes of URL, title, format, holdings, or publication status of e-journals. Automatic URL checking was considered the responsibility of the systems office. Hudson and Windsor (1998) reported on the types of access provided by the Ohio University Libraries at that point. The Libraries used one bibliographic record to provide access to both the electronic and the print versions of a journal title; they also listed electronic journals alphabetically on their web site. In addition, the Libraries encouraged subject bibliographers to develop their own web pages listing electronic serials in their corresponding subject areas. Knudson et al. (1999) described the Research Library at Los Alamos National Laboratory's comprehensive way of generating web-page access to e-journals by using both standard and local data in MARC records in their OPAC. Simple title-word searching capabilities were added to their e-journal web site in response to user requests. They began adding indexing information to the records in both their OPAC and their web site for e-journals for which they had created article-level links from local citation databases. Given our current tools, identifying and collocating the e-journals in a particular subject area that were cataloged in a library's OPAC can be problematic. Shadle (1999) outlined the difficulty of using the Library of Congress Subject Headings (LCSH) for such retrieval; LCSH, unlike the Medical Subject Headings, does not have the form subdivision "--Electronic journals." He surveyed 108 libraries and summarized the various ways that libraries are using other elements of the bibliographic record to help users isolate e-journals in their OPACs. Davis (1998) called for a reevaluation of the definition and application of the form subdivision "--Periodicals" because of the increasing number of e-journals being cataloged. Schaffner (1998) described how the Brandeis Science Library's web page for e-journals was organized (by broad subject area and alphabetically by title) and expressed her belief in the usefulness and desirability of broad subject listings, whether they are generated from the OPAC or created independently. Fosmire and Young (2000) surveyed the level of access provided to selected, free scholarly electronic journals by academic libraries, and found that basic information about these online journals was only provided by between 50 and 66 percent of the institutions, despite the fact that approximately half of the titles were indexed by major indexing and abstracting services, and that some of the journals were no longer published in print. The difficulty of providing traditional means of access, such as national bibliographies, to resources as volatile as e-journals is demonstrated by Langballe (1999), who described the results of attempting to access 299 Internet serials listed in the Norwegian national bibliographic database Norper as of November 1997. Seventy-four titles were found neither at the original site nor at a new one; fifty-three of the found serials had probably ceased. Langballe stressed that link checking could not be left solely to electronic programs and search robots; regular checking by library staff was also required. Links to e-journal articles: a further level of access Moothart (1996a) outlined the capabilities of Journals Online, a system developed by Bath Information and Data Services that links a web-based bibliographical database to e-journal articles. Although the articles were loaded on publishers' servers, the user interacted with only one interface. Moothart pointed out that such a system cannot obviate the collection-development role of librarians, who select out of a number of options the best resources for their users. Hewett (1997) reported on a conference (held at the University of Warwick in 1997) at which the experiences and findings of the DECOMATE (DElivery of COpyright MATerials to End-Users) project, funded by the European Commission, were discussed. The principal objectives of the project were to design and implement a generic electronic document-delivery system that would enable users to search various bibliographic databases and retrieve the full text of articles. At the same conference, the subscription agents Swets and the publisher Elsevier Science presented similar, if non-generic, systems, that they were developing. Aggregator databases Coming onto the market in the late 1990's and dubbed "aggregator databases," large databases (similar to those developed by Swets and Elsevier) such as EBSCOHost, Lexis-Nexis, and ABI/Inform presented librarians with many complex choices regarding bibliographic control of those e-journals accessed thereby. Riemer (1999a) pointed out that for many libraries the primary benefit of aggregator databases is that they can provide affordable access to many serial titles not previously held. The primary drawback of such databases for libraries stems from the frequent changes by vendors in content (e.g., titles added and dropped) and inconsistencies in type of coverage (e.g., abstracts versus full text). These problems are compounded by inadequate communication from the vendors to libraries about these changes and inconsistencies. Dennis (1999) outlined his experiences with such difficulties as a reference/collection-development librarian and yet still called for "fully analyzed, title-by-title cataloging, and detailed holdings, for each and every title in each and every aggregator database to which we offer access." Riemer (1999b) reported on a survey conducted in December 1998 by the Program for Cooperative Cataloging Standing Committee on Automation Task Group on Journals in Aggregator Databases on how CONSER members were providing access to journal titles within aggregator databases. While the second most frequently used method of providing access was by means of listing titles on web sites, far fewer respondents actually preferred this method; 71 percent of the respondents wanted some sort of bibliographic record in the OPAC to represent e-journals. Perhaps the increased collaboration that Riemer and Dennis called for, both among librarians, and between vendors and librarians, in addition to the development of new technologies, will enable librarians to provide the desired bibliographic access to these resources. Methods Since virtually no research had been done on this subject at the time that this study was conceived, the research team, which consisted of both catalogers and reference librarians, drew upon the practices of their institutions to create a list of science and technology e-journal access methods: bibliographic records which describe only the electronic versions of journals (called the "separate record approach" in CONSER documentation (Hirons 1999)); bibliographic records where data for the electronic version of a title has been added to the record for the print version (called the "single record approach" in CONSER documentation (Hirons 1999)); web-based HTML lists of e-journal titles; aggregator web sites, which provide access to a package of e-journals; and database-driven applications which build dynamic HTML lists of e-journal titles. For this study, a four-part questionnaire was designed. It consisted of 42 questions that were a combination of open-ended questions and forced-choice items. When appropriate, forced-choice items included an "other" category to obtain answers that were not included in the multiple-choice set. The goal of the first part of the questionnaire was to create a profile of the respondent's library, including type and focus of institution, student enrollment, size and format (web, Windows, text) of catalog, and library staffing.1 Given the limited number of colleges and universities specializing in science and technology, survey respondents could come from science and technology, liberal arts, or comprehensive schools. The questions posed in the remaining three sections of the questionnaire were formulated by librarians who currently serve in public and/or technical services roles in science and technology libraries. The second part of the questionnaire addressed e-journals in science and technology subjects in the respondent's collection. Details such as publishers, methods of purchase, and connection methods (IP filtering, logon/password, etc.) were gathered. The third part revolved around e-journal access methods, the nature of those access methods, and maintenance issues regarding both changes in content coverage and changes in URLs. For example, if access points for the e-journal were added to the bibliographic record for the print version, were these access points active links, such as hyperlinked text or images, or were they unlinked notes directing users to URLs or other methods of access? The last part of the questionnaire dealt with the respondent's personal reaction to the success(es) or failure(s) of each method of e-journal access as implemented at his or her institution; that is, what the respondent encountered while working the desk and helping patrons, and how respondents (not the patrons themselves) felt about the access methods experienced. The research team hoped that since reference librarians have ample opportunity to assist both users and staff members in accessing e-journals, that they, as a group, could provide data describing their library experiences that would be useful in shaping future choices made by cataloging, reference, and computer-services staff regarding e-journal access. Are e-journal cataloging projects a waste of time? Web page listings of e-journals are commonplace, but how are they being generated and maintained? A draft version of the questionnaire was tested by some team members' public-services colleagues in their home institutions for clarity and appropriateness. Its final version was then posted on the web site of a team member's library as a web form. A request for participation was solicited through a number of professional lists relating to science and technology librarianship, reference, and cataloging (cataloging was included because some catalogers also work as reference librarians). Data were collected electronically, through form submission. Completed forms were collated and the results tallied. The data gathered were calculated into percentages and compared. The findings from this sample were then inferred to the profession at large. (A text version of the questionnaire is included as an appendix at the end of this paper.) Results Part 1. Sixty-two filled out questionnaires were returned to the Committee, out of which only sixty were eligible for analysis. The remaining two questionnaires were disqualified due to being incomplete. Respondents represented institutions which ranged from two-year colleges to research universities, including both science and technology institutions and comprehensive institutions. Enrollments at the respondents' institutions ranged from fewer than 5,000 students to over 40,000 (Figure 1: A, C). Figure 1. Distribution of respondents by institution type (A, B) and enrollment (C) The institutions of all respondents had a web site. Fifty-two (86.6%) respondents worked at institutions which had a web-based catalog interface; six (10%) had a text-based interface; and two (3.3%) had a GUI (Windows) interface. The number of bibliographic records in their catalogs ranged from 30,000 to over ten million. The respondents themselves were a mixed group. Approximately one-third worked in general (all subjects) libraries; one-third in science and technology branch libraries; and one-third in science and technology libraries (these last two distinctions caused some confusion among the respondents and were likely the result of lack of explanation on the survey itself; certainly, only seven respondents (Figure 1: B) described themselves as working at science and technology institutions). Respondents came primarily from public services, as expected, but technical services and other areas were represented as well. Most respondents worked as reference librarians in science and technology libraries, or branches dealing exclusively with science and engineering subjects; some were assigned science and technology reference duties within a general library environment, and a few fielded reference questions on all subjects. The respondents who worked as catalogers primarily catalog all subjects and all formats without any specialization, although a few were assigned particular subjects or formats as shown in Table 1. Table 1. Characteristics of respondents & their organizational setups. Public Services 36 Both Technical and Public Services 17 Other 7   Reference: All subjects 8 Reference: Assigned sci/tech 15 Reference: Sci/tech library 37   Catalogers: All subjects/formats 41 Catalogers: Assigned subjects 8 Catalogers: Assigned formats 8 No reply 3 Part 2. The institutions represented by the respondents make available to their users anywhere from twenty to "thousands" of e-journals in the sciences and engineering. The institutions of all sixty respondents subscribe to journals by both commercial publishers, such as Wiley or Elsevier, and professional societies, such as the American Chemical Society or the Institute of Physics. This would be expected given the prominence of both commercial publishers and scientific societies in scientific communication. The institutions of a smaller number of respondents (46, or 76.6%) receive some e-journals from special projects such as SPARC (Scholarly Publishing & Academic Resource Coalition) or JSTOR (Journal Storage: The Scholarly Journal Archive). Paid subscriptions are obtained directly from the publisher (49 respondents, 81.6%); through a subscription service such as EBSCO (43 respondents, 71.6%); through an aggregator (47 respondents, 78.3%); or via consortial purchase (49 respondents, 81.6%). Connections to e-journals are supplied through various means, such as IP limiting, login/password combinations, and dedicated terminals. Campus-wide IP limiting was by far the most common method used by the respondents' institutions (60 respondents, or 100%). The next most common methods of access were a user typing a login/password (27 respondents, 45.0%) and an embedded login script that would be transparent to the user (21 respondents, 35.0%). A much smaller number of respondents work at institutions which employ library-wide IP limiting (11 respondents, 18.3%), login required by librarians on behalf of patrons (7 respondents, 11.6%) or dedicated terminals (2 respondents, 3.3%), Part 3. The five possible access methods for e-journals explored by the survey are outlined in the Introduction. The first approach to e-journal access (and one that was the most labor-intensive for technical services departments prior to when vendor-supplied MARC records were made available) is to create separate records for e-journal titles, as many catalogers believed to be appropriate under AACR2r. A significant number of the respondents (46, 76.6%) reported that their institutions use this access method for at least some of their e-journals. A similar number of respondents (44, 73.3%) reported that their institutions use the single record approach by adding access information about the electronic version of a journal to the bibliographic record representing the paper version. Although not explicitly included in AACR2r, this method of providing access has been deemed an acceptable option by the CONSER Program (Hirons 1999), which creates a large portion of bibliographic records for serials available in national bibliographic databases. This method of access was popular because it took the least amount of time and effort to make the e-journals accessible through the OPAC by simply adding a URL to existing records without extensive record review. One of the earliest methods of e-journal access, and one that remains common, is the HTML list containing hyperlinks to e-journal titles. Librarians faced with the need to provide quick access to a new medium usually constructed these lists as a short-term solution; the lists remain with us even with the increasing number of bibliographic records in online catalogs. Lists generally appear in three manifestations: arranged by subject, arranged by title, and arranged by aggregator.2 The institutions of 32 (53.3%) respondents maintained subject lists, those of 40 (66.6%) respondents maintained title lists, and those of 26 (43.3%) maintained aggregator/vendor lists. Responsibility for maintaining the lists was spread across library departments. In addition to HTML lists by aggregator or vendor, the institutions of over half of the respondents (34, 56.6%) provide a connection directly to an aggregator's or vendor's general web site, instead of to specific journals within that web site; that is, to the American Institute of Physics site, instead of the top page for Physical Review A. Once within that web site, users may be limited in what they can access depending on institutional subscriptions. The e-journal access method in its earliest stage of development is the use of database-driven applications that use a search query to generate dynamic HTML lists. Twenty-three respondents' institutions (38.3%) are experimenting with this access method. Maintenance responsibilities rested primarily with library computer services staff. In addition to these common methods of e-journal access, three respondents indicated that they also use other methods, but did not specify. When assigned, the responsibility for checking and/or updating URLs is distributed much as one might expect: according to the access method. Technical services or acquisitions departments are generally responsible for receiving and/or checking URLs in bibliographic records; public services departments are generally responsible for receiving and/or checking URLs in web-based lists, and computer services departments are generally responsible for receiving and/or checking URLs for database-driven applications. However, URL checking is an assigned practice at only a small number of the respondents' institutions. Most rely either on users or staff to alert them to access problems, or use other methods such as automated link-checking programs. Part 4. The core of the survey was data collection on the perceived successes and failures of these e-journal access methods. The Committee did not intend that respondents attempt to provide empirical data about usage -- that and similar questions, such as user responses to various access methods, were beyond the scope of this survey. Rather, we hoped that reference librarians would provide data about their personal experiences on the desk as they assisted both users and staff. As shown in Figure 2, the median response to the query on the methods used for e-journal access was 68%, with the least response rate for the database-driven applications method. Among respondents for the database-driven applications method, opinion about its preference was equally divided. Similar was the case with the web-lists method. Extremely low preference rates (<20%) to aggregator sites and the separate-record access methods were noted among the respondents. Figure 2. Comparison of Response rate and level of preference for various e-journal access methods The first access method, creating separate records for e-journals, was one of the access methods at 41 (68.3%) of the respondents' institutions. Of these 41 respondents, 8 (19.5%) chose it as their preferred access method; 33 (80.5%) chose "not preferred."3 Perceived strengths of this access method include: Integration into the online catalog allows "one-stop shopping" for users (34, 82.9%) Access points and bibliographic data specific to the online version of the journal (23, 56.1%) Consistency in access across the collection (19, 46.3%) Coverage clearly indicated for the electronic version (16, 39.0%) Conformance to national cataloging standards (13, 31.7%) Availability of general bibliographic data (11, 26.8%) Other (2, 4.9%). The perceived weaknesses of separate bibliographic records for science and technology e-journals include: Users sometimes don't find the record for the electronic version, leading them to assume we don't have it (27, 65.9%) Changes in the bibliographic records (for URLs, subscription changes, back files) don't appear quickly enough (21, 51.2%) Users don't like having to search the catalog to get to the active link (19, 46.3%) Cataloging journals in this manner takes more time (18, 43.9%) Staff time devoted to necessary maintenance is too great (15, 36.6%) In our catalog, it's difficult to isolate a group of these records by subject (12, 27.3%) Other (4, 9.8%) The next access method, the single record approach, is used at 44 (73.3%) of the respondents' institutions. Of these 44 respondents, 16 (36.4%) indicated this as a preferred method of access; 24 (54.5%) chose "not preferred," with four (9.8%) not responding to the question. Perceived strengths include: Access points for both titles (38, 86.4%) Integration into the online catalog allows "one-stop shopping" for users (37, 84.1%) Current subscriptions appear on the same record, allowing users to see the availability of each (32, 72.7%) Availability of general bibliographic data (20, 45.5%) Other (12, 27.3%). Perceived weaknesses of bibliographic records containing information for both paper and electronic versions of a journal include: Variations in paper vs. electronic "holdings" lead to confusion for staff and users (26, 59.1%) Changes in paper and electronic versions of a journal may not be equivalent (20, 45.5%) Users don't like having to search the catalog to get to the active link (17, 38.6%) Changes in the bibliographic records (for URLs, subscription changes, back files) don't appear quickly enough (17, 38.6%) Users sometimes don't find the link for the electronic version, leading them to assume that we don't have the journal (15, 34.1%) In our catalog, it's difficult to isolate a group of these records by subject (10, 22.7%) Staff time devoted to necessary maintenance is too great (8, 18.2%) Other (3, 6.8%). Web lists, whether by subject, title, or aggregator/vendor, were one of the e-journal access methods at 48 (80.0%) of the respondents' institutions. Of those 48 respondents, an equal number (22, 45.8% in each case) chose this as their "preferred" or "not preferred" access method; 4 respondents (8.3%) did not answer this question. Perceived strengths include: Easy to see holdings by title (42, 87.5%) Users find lists easier to use than the catalog (32, 66.7%) Staff find lists easier to use than the catalog (24, 50%) Easy to see holdings by subject (24, 50%) Easy to see holdings by vendor (9, 18.8%) Other (3, 6.3%). Perceived weaknesses of web lists include: Large lists are difficult to manage and update (31, 64.6%) Full bibliographic access is not available (21, 43.8%) Time devoted to necessary maintenance is too great (20, 41.7%) Lists keep users from learning how to use the catalog effectively (19, 39.6%) Users won't let us drop the lists (8, 16.7%) Other (4, 8.3%). The use of aggregator sites was one of the access methods used at 40 (66.6%) of the respondents' institutions, and the only e-journal access method at one institution. Of those 40 respondents, 36 (90%) marked "not preferred" for this access method; none saw it as a preferred method. Perceived strengths include: Access to journals to which the library does not normally subscribe (27, 67.5%) All or a large group of the library's e-journals are in "one place" (13, 32.5%) Users find a single site easier to use than the catalog (9, 22.5%) Staff find a single site easier to use than the catalog (9, 22.5%) It requires maintaining only one URL (8, 20%) Other (2, 5.0%). Perceived weaknesses of aggregator/vendor web sites include: When the site is down, e-journal access is cut off (20, 50.0%) Subject coverage is inadequate for our purposes (17, 42.5%) No full bibliographic access (16, 40%) No good access by subject (12, 30%) Other (9, 22.5%). Of the 23 (38.3%) respondents whose institutions use database-driven applications for e-journal access, 11 (47.8%) saw it as a preferred method; 12 (52.2%) chose "not preferred." Perceived strengths of this method include: Easy update of URLs and other data (22, 95.7%) Database permits storage and retrieval of other information pertinent to managing e-collection (21, 91.3%) Other (3, 13.0%) Staff time devoted to maintenance is minimal (1, 4.3%). Eight surveys had responses about the perceived weaknesses of the database-driven applications method, but only seven respondents worked at institutions where this method is used. Comments about the perceived weaknesses of database-driven applications from the seven surveys fell into the "other" category where respondents described a wide range of perceived weaknesses in their own words. Discussion Part 1. This study has several limitations. First, by placing the survey instrument on the web and soliciting replies over various Internet library-related lists without doing sampling of our survey population first, the response rate cannot be measured, nor can it be guaranteed that there was only one respondent per institution. To eliminate such drawbacks, if we were to conduct this survey again, we would rather do a sampling and use mixed-mode surveys to solicit replies. Second, the survey instrument itself was probably overly long; although the demographic and procedural data are interesting, they have proven unnecessary for the current analysis. Certainly occasional contradictions within some of the responses indicate that some questions may have been inadequately defined. If we had conducted a pilot survey, we might have realized that we needed to shorten the survey to avoid a "fatigue factor" or needed to clarify certain terms. Third, the authors have little if any experience with constructing and conducting survey research, and as a result may have made errors in the plan or execution of the study. Fourth, our survey failed to ask respondents whether they preferred any combination of access methods. A 1998 survey of the library community conducted by the Program for Cooperative Cataloging Standing Committee on Automation Task Group on Journals in Aggregator Databases investigated opinions about the most desired form of access to journals in aggregators; the Task Group found that "half of the people who chose 'none of the above' explained that they did so because the survey required a single answer, and they used the comment section to indicate they wanted a combination of two or more of the first three choices" (Riemer 1999b). Before any one method of access is determined to be undesirable, the desirability of combinations of methods, especially those that can be generated automatically from the same source data, should be evaluated. Part 2. The connection methods used for science and technology e-journals at the respondents' institutions would seem to be very indicative of what librarians want for their patrons in terms of optimal convenience and accessibility while still retaining the control over the resource that a publisher would expect, with campus-wide IP limiting being the favored choice by a wide margin; this method allows the most desired benefit of e-journals -- service across the campus community -- while preserving the standard method of access for those not affiliated with the campus, that of having to come into the library to access the journal. That use agreements still sometimes demand that e-journal use be limited only to people affiliated with the campus indicates that some publishers are still thinking of their e-journal products both as "special" and in terms of their development costs, not as the next step in an evolutionary process of the development of scientific communication. Fewer than 20% of the respondents work at institutions that use library-wide IP limiting; limiting the ability to access an e-journal to within the physical walls of a library eliminates one of the chief potential benefits of its electronic form: campus-wide accessibility. Logins, whether physically typed in or embedded within scripts, and dedicated terminals tend also to elide the access and convenience that e-journals are meant to provide, and their dramatically lower incidence of usage as indicated by the respondents suggest that these might be accepted by librarians only when there is no other alternative (i.e., the publisher cannot be convinced that campus-wide IP limiting makes sense or claims that this method of connection is not available). Librarians should continue to work with publishers to promote the best possible access for their patrons. Part 3. Five access methods were explored by this survey. The first two methods, both of which involve bibliographic records in the library's catalog, were far more common than any of the others except for HTML lists by title. At this point in the history of the science and technology e-journal, institutions are adding records to their catalogs for e-journals themselves or adding connection information (in either hot linked or note form) to the bibliographic record for the paper version of the title. Both methods are used in some catalogs. That the more labor-intensive separate-record approach to cataloging is used at the institutions of nearly 70% of the respondents may indicate that e-journals are seen as a vital part of the collection, to be put on a footing with more traditional library materials. Some institutions may be choosing this approach because it allows them to extract records from their catalogs more easily once they no longer need a link to a particular e-journal. In addition, e-journal subscriptions are becoming more stable, with fewer short-term or trial subscriptions being offered. Adding e-journal information to the record for the paper version of the journal may have begun at some institutions as a stopgap measure, a temporary solution until full records could be added to the catalog. At other institutions, however, this policy is standard, possibly because of limited cataloging resources, or possibly because of choices regarding approaches to reference practice and service to patrons. In the case of at least one institution, this policy was adopted as a result of a cost-effectiveness study that considered those issues as well as the availability of records, either from vendors or a bibliographic utility such as OCLC, describing the electronic version of journals. Flat HTML lists, whether arranged by title, subject, or aggregator/vendor, were one of the earliest science and technology e-journal access methods, a solution that could be implemented and updated by public services staff while subscriptions stabilized and the catalogers were brought on board. These lists serve as kind of a cross between an index and a guide: they permit alphabetical or subject grouping, and have the additional property of being "searchable" through the Find function on browser software. Even with the significant number of bibliographic records in library catalogs, the continuing presence of HTML lists indicates that they remain both useful and usable. Over half of the respondents work at institutions which provide some kind of link directly to a vendor's or aggregator's web site. While contracts with aggregators provide some acknowledged benefits, a user who is simply "deposited" at an aggregator's page will have no guidance as to which e-journals are actually accessible, and possibly have little feeling about how to navigate the site or find titles applicable to the user's field of study. Users who find themselves in this situation are probably not likely to think positively of the library. The newest and least common method of e-journal access explored in this study -- that of database-driven applications (DDAs) used to dynamically generate HTML lists -- is clearly in a developmental stage. The variety of platforms and system software, scripting software, and maintenance/data update methods indicate that DDAs are an "emerging technology" in terms of managing e-journal access. It will be interesting to see if this access method continues to grow in popularity, and if, for example, PERL maintains its popularity as a scripting language. Also, DDAs (such as those developed at the Research Library at Los Alamos National Laboratory) which generate HTML data for web sites from MARC records (i.e., using the catalog as their data source), may become more widely used, because they minimize the labor-intensive task of maintaining parallel information about e-journals in both library catalogs and web pages (Knudson et al. 1999). Part 4. The core of this study attempts to gather data regarding what reference librarians have observed during their time on the desk about these various science and technology e-journal access methods. Such data are necessarily subjective and anecdotal to a certain extent, but regular, reflective practitioners provide crucial information that can assist all involved parties in designing and implementing more effective, efficient, user-friendly systems. With new catalog technologies on the horizon, these areas are ripe for exploration. Separate bibliographic records for science and technology e-journals are used at almost 70% of the respondents' institutions, but more than 80% of those same respondents do NOT prefer this as an access method. The most significant benefit by far was seen to be integration into the catalog, which cuts down on the number of places a user must search. This feeling probably illustrates a general desire throughout the profession to get everything -- from bibliographic records to citation databases to full text and beyond--into one information system. From a technology standpoint, this day is approaching. Respondents also liked access to item-specific bibliographic data and consistency in access across the collection, which certainly makes the library seem less capricious. These positives seem to be overridden, however, by users' difficulties in actually finding the bibliographic record in the first place, by record updates not happening quickly enough, and by users not wanting to search the catalog to find the hot link. The first of these difficulties could be addressed to some extent in bibliographic instruction sessions, although not without difficulty. Addressing the second requires positive relationships and cooperation between public and technical services departments, but if a large number of records need to be updated, that will involve a time lag even in an efficient environment. The last of these three major objections is understandable and may be beyond resolution. Adding information about an e-journal to the bibliographic record for its paper version is an access method used at the institutions of almost 75% of the respondents, and over half of those do NOT prefer this as an access method. Positive aspects of this approach for many respondents include the availability of access points for both titles on one record; the integration of more items into the catalog; and the appearance of subscription information for both titles on one record, which allows users to see the availability of both versions. Again, we see here a trend toward a desire for a comprehensive database/catalog. Even though librarianship may have developed different tools for different reasons, such as catalogs for access to books and journal titles, and indexes for access to journal articles, new technologies, whether software-based or not, are beginning to provide a unified access platform. The two major objections to this access method are grounded in integration itself: differences in holdings can be confusing for users and staff, and changes to the journal may not be equivalent in both the paper and electronic versions. The inclusion of both sets of information on one record implies an equivalence when there may not be one. This approach has certain virtues, but it contradicts what we try to teach users about the content of the catalog and undermines their experience of consistency throughout the rest of the catalog database. An equal number of respondents prefer or do NOT prefer flat HTML lists as an access method to science and technology e-journals. The greatest strengths of these web-based lists include ease of seeing holdings by title and ease of use for both users and staff. And these lists are easy to use; who wants to search the catalog when it's possible to just click on a couple of links? Even given the major perceived weaknesses of this access method, such as lists being difficult to update and manage (though not, apparently, in a time sense; only about 40% of those respondents whose institutions use such lists objected to them on grounds of lists being a time sink), and the tendency of lists to undermine users' learning to use the catalog effectively, the neat division of preference indicates that "ease of use" is indeed a driving force in access choices. Bibliographic records, the cornerstone of the library catalog, are somehow beyond "ease of use." Aggregator/vendor web sites are used at 66% of the respondents' institutions, and are overwhelmingly disliked. The only significant benefit from aggregators, according to two-thirds of the respondents, is that through aggregators a library can get access to journals to which it does not normally subscribe. The major objection to this access method is that a down connection or site means that e-journal access is cut off, with no recourse. The greater the library's dependence on an aggregator for e-journals, of course, the greater the problem. This study did not delve into aggregator/vendor web site issues in greater detail; this area might support further investigation. Database-driven applications were the other access method besides HTML lists where respondents split in their evaluations of it as a preferred method. Overwhelmingly, ease of update and data storage ability were seen as the major strengths. No single weakness predominated. Conclusion and Directions for Future Research The results of this survey support our assumptions that most institutions provide more than one method to access their full-text e-journals, and the responsibility for managing these access methods is widely distributed throughout the library among staff in public services, technical services and systems/information technology divisions. Generally speaking, librarians are not totally satisfied with any of the five methods being surveyed, although the web-based listing seems to have a slight popularity over the others for its ease of use and its capability to group e-journals by subjects. It is also the most used method. Although less than half of the respondents' institutions provide access through some database-driven, dynamically generated web lists, this method has the highest percentage of "preferred" votes among its users. Using catalog records to provide access to e-journals, be it the single-record approach or the separate-record approach, is not preferred by our respondents. Despite the fact that catalog records usually contain fuller, more accurate bibliographic and access information about e-journals, there is tremendous concern about the users not being able to locate the active links in the already confusing catalog, the delay in updating changes, and the lack of grouping like-materials by subjects in most integrated library systems. Before we advocate abandoning traditional cataloging for e-journals, an investigation into how much of a user's confusion stems from the bibliographic records themselves and how much from a particular OPAC's interface seems crucial. No respondent prefers single linking to the aggregator's web site, including the one respondent's institution where this is the only access method provided. Even though it is easier to use a single web site and maintain a single URL, librarians do no like to have their service pending on the availability of the vendor's web site, which usually does not provide good subject access to their e-journals. The survey results demonstrate that librarians consider user convenience, clarity, ease of use, timely update, subject access, and minimal maintenance as the criteria for effective access method to e-journals. Several months have passed since the survey was conducted. During this time, new technological and standard developments have emerged, resulting in new services and products as attractive options for libraries looking for ways to improve access methods to e-journals. For example, some ILS vendors have adopted Open Linking technology based on OpenURL standards, allowing direct linking of citations from indexes and abstracts to full-text articles. E-journal management vendors such as SerialsSolutions, TDNet and JournalList are offering products and services such as MARC-like records, and HTML/XML based files of customizable data about an institution's e-journal subscriptions, with up-to-date URL and coverage, enabling libraries to provide multiple access methods, through the catalog and/or the web listings, without having to maintain dual files. Further research that performs cost/benefit analyses of various access methods or combinations of methods would enable librarians to make better use of their resources and might spur technological developments that would reduce costs. Before library administration pours over more money for the Next Great Thing, further research on access issues from the users' perspective would be helpful in defining the true benefits of these and other access methods, bridging the gap between perception and reality. A new environmental scan and a well designed usability study for faculty, students and researchers can help to achieve this goal. Notes 1 The intervals for student enrollment used in the survey were devised without consideration of any standard intervals used by an organization such as the National Center for Education Statistics. If we were to conduct a similar survey, we would use a standard set of intervals. 2 We did not include in our survey a question about lists of e-journals arranged by the learned societies that issue them. 3 When indicating perceived strengths and weaknesses of particular access methods, respondents were allowed to select more than one response, so percentages do not add up to 100%. References ACRL/STS Subject and Bibliographic Access Committee. 1999. Electronic publishing of scholarly journals: a bibliographic essay of current issues. Issues in Science and Technology Librarianship. [Online]. Available: http://www.istl.org/99-spring/article4.html [August 8, 2002]. Davis, C.N. 1998. Problems in defining '-periodicals' (electronic or otherwise). Serials Review 24(2):107-109. Dennis, S.L. 1999. Aggravating or aggregating? providing effective access to contents of aggregator databases: a reference/collection development librarian's perspective. Cataloging & Classification Quarterly 28(4):15-30. Fosmire, M. & Young, E. 2000. Free scholarly electronic journals: what access do college and university libraries provide? College & Research Libraries 61(6):500-508. Foster, C.L. 1998. Through the arch: electronic serials from acquisition to access. Library Acquisitions 22(4):471-474. Hewett, S. 1997. Journals in transition: from paper to electronic access: the European Union project DECOMATE: a case study. Serials Review 23(2):79-83. Hirons, J.L., (ed.). 1999. CONSER Cataloging Manual. Section 31.3.5. Hruska, M. 1995. Remote Internet serials in the OPAC? Serials Review. 21: 68-70. Hudson, L. & Windsor, L. 1998. Providing access to electronic journals: the Ohio University experience. Against the Grain 10(3):1, 16, 18. Jones, W. 1995. We need those e-serial records. Serials Review 21:74-75. Knudson, F.L. et al. 1999. Leveraging the MARC record: automatic generation of electronic journal web pages. Serials Review 25:1-11. Langballe, A.M.H. 1999. Remote access electronic serials and the National Library of Norway. International Cataloguing and Bibliographic Control 28(3):78-79. Moothart, T. 1996a. Journals Online expands the options for e-journal search and retrieval. Serials Review 22:127-130. Moothart, T. 1996b. Providing access to e-journals through library home pages. Serials Review 22:71-77. Moothart, T. 1997. Muddy waters bring the library blues and enhanced electronic access to journals. Serials Review 23(4):79-82. Morgan, E.L. 1995. Adding Internet resources to our OPACS. Serials Review 21:70-72. Reynolds, R. 1995. Inventory list or information gateway? the role of the catalog in the digital age. Serials Review 21:75-77. Rich, L.A. & Rabine, J.L. 1999. How libraries are providing access to electronic serials: a survey of academic library web sites. Serials Review 25(2):35-46. Riemer, J.J. 1999a. Aggregators -- new challenges to bibliographic control. Cataloging & Classification Quarterly 28(4):5-6. Riemer, J.J. 1999b. CONSER's aggregator survey and the work of the PCC Task Group. Cataloging & Classification Quarterly 28(4):7-13. Rodgers, D.R. 1999. Eeee!-serials: providing access to online serials. Serials Librarian 36(3/4):467-473. Schaffner, A. 1998. Lost in cyberspace? subject access to electronic journals. Serials Review 24(2):102-104. Shadle, S. 1998. Identification of electronic journals in the online catalog. Serials Review 24(2):104-107. Simser, C.N. 1998. Through the arch: electronic serials from acquisitions to access. Serials Review 24(3/4):144-147. Sleeman, A.M. 1995. Cataloging remote access electronic materials. Serials Review 21:72-74. Woodward, H. & McKnight, C. 1995. Electronic journals: issues of access and bibliographical control. Serials Review 21:71-78. Appendix View the "Questionnaire for Perceived Successes and Failures of Sci/Tech E-Journal Access: A Comparative Study". This is a 600K Word document. Previous   Contents   Next