How people use electronic resources or their preferences for print and electronic library services have been the focus of dozens of individual research studies in the last few years. (In this report, these studies are designated as Tier 2 studies to differentiate them from the major, ongoing studies reported as Tier 1.) Surveys are by far the most popular method, with academic faculty and students the most popular participants. Although participants include a variety of subject experts, scientists of various sorts have been studied most frequently. Not every study will be discussed in detail in this analysis, but both common threads and unique findings of the studies will be highlighted. Summary analyses of the methods used and groups studied are shown in Figures 3.1, 3.2, and 3.3.


The Tier 2 studies echo many of the findings of the larger Tier 1 studies, allowing some consistent conclusions to be drawn about user behavior. This section is arranged by general themes that are found in many studies, and then further organized by specific themes, some of which emerge in only some of the studies, others that emerge consistently, but with contradictory results. The general themes and specific sub-themes are as follows:

  1. Differences in behavior or preferences that can be explained by differences among users. Differences include :
    • differences by subject discipline
    • differences by user status or workplace
    • differences by task
    • differences by age or gender
  2. Information-seeking behavior and preferences, including differences between print and electronic resources:
    • browsing versus searching
    • preferences for print or electronic resources
    • awareness of electronic resources
    • search strategies
    • reasons for using the resources
    • sources of information about resources
    • self-evaluation of system navigation
  3. Perceived advantages of electronic resources and preferences, including:
    • how electronic resources improve workflow or save time
    • preferred features of electronic information systems
    • currency and timeliness of sources
  4. Perceived disadvantages or concerns about electronic resources, including:
    • technological or service problems
    • archiving
    • problems or confusion with information systems
    • preferred formats for reading
    • electronic versus print resources
  5. Library policy and financial issues:
    • willingness to pay for electronic information
    • willingness to cancel print journals in favor of electronic
    • other library budgetary issues that affect users

3.1 Differences in Behavior or Preferences that can be Explained by Differences Among Users

The concept of a single typical “user” of information systems is clearly a fallacy. Scientists seek and use information differently than do social scientists or humanists; undergraduate students behave differently than do graduate students or faculty; searching for information for personal use is different from searching for work-related tasks. Numerous studies have reached these conclusions.

Faculty members and other professionals in the sciences, math, and medicine fields were early adopters of electronic journals and other digital library resources and remain the heaviest and most enthusiastic users (Kidd 2002; Tenner and Yang 1999; Voorbij 1999; Hiller 2002; Rowley 2001; Dillon and Hahn 2002). Lenares (1999) found as early as the late 1990s that 90% of the physical science faculty used electronic journals at least part of the time, compared with 61% of all faculty users in ARL universities. The percentage of faculty using electronic journals for at least some of their readings increased from 1998 to 1999 and has continued to increase each year since (Lenares 1999). By 2002 at the University of Maryland, more faculty members used electronic journals daily or weekly than they did print journals (Dillon and Hahn 2002). This corresponds to the decrease in physical visits to the library by graduate students and faculty, especially in health sciences, science, and engineering (Hiller 2002). (This is discussed in more detail in section 3.5).

Enthusiasm for electronic journals and patterns of use vary even among fields of science. Chemists and physicists use them frequently, while earth scientists and mathematicians see fewer advantages (Mahe, Andrys, and Chartron 2000). Medical school users rely on fewer electronic journal titles for their downloads and readings than do other library users (Davis 2002). These variations are also reported by Tenopir and King, who found physicists and astronomers to be among the most enthusiastic users of electronic articles, partly because the digital e-print archives (, the Astrophysics Data System (ADS), and e-journals of the American Astronomical Society were designed specifically to facilitate their natural work patterns.

Business school faculty members were also early adopters (Tomney and Burton 1998). Business school faculty reported the highest use, while Palmer and Sandler (2003) found economics faculty to be the most enthusiastic users of electronic journals. Speier et al. (1999) and Hahn et al. (1999) found that among business faculty, finance and management information systems faculty were more aware of electronic journals than those in other fields. Faculty members in history, education, and the arts have been slower to adopt electronic journals (Tomney and Burton 1998). Among corporate users, investment and banking companies or departments spend a higher percent of their budgets on electronic products than do other types of businesses such as pharmaceutical firms, legal services, food services, or telecommunications (Carrick 2002).

Although high percentages of faculty members use electronic journals, they still use a variety of sources, including print, for their readings (Tenopir and King; Brown 1999). In 1999, Lenares found that although 90% of physical sciences faculty respondents at ARL libraries used electronic journals, half reported that they read articles from electronic journals infrequently.

Undergraduates in the life science disciplines were found by Whitmire (2002) to engage in more information-seeking activities than were students in other disciplines, including using the online catalog, asking librarians for help, using indexes, and browsing the stacks. Engineers engaged in the fewest information-seeking behaviors.

Variations in information-seeking behavior by individual groups make the task of designing an e-journal, user interface, or electronic library a challenge. One attempt to meet this challenge is to provide a “MyLibrary” feature on a library’s home page. This allows users to customize their view of the electronic library and highlight the resources they use most often (Ghaphery 2002).

Another solution is to maintain a balance between print and electronic collections, depending on the preferences of the main user groups. (This will be discussed in more detail in sections 3.2 and 3.5). Librarians must gauge their systems, collection development decisions, and instruction with user differences firmly in mind. No single solution will be best for everyone and the totally digital library is a long way in the future in many subject disciplines. Publishers also must design electronic resources that facilitate the work patterns of their target audiences.

Differences in motivation or task also cause variations in information seeking and use. Nelson (2001), in a study of faculty and students at the University of West England, found that the greatest predictor of electronic use was whether or not the person was engaged in research. Researchers and academic staff were more likely to use electronic journals than were administrative staff. King and Montgomery (2002) found that more than half of the readings by faculty and doctoral students are done while conducting primary research. Faculty who publish more and those who had served on promotion and tenure committees were found to be more likely to be aware of electronic journals, and faculty with tenure are more likely to submit articles to electronic journals (Hahn et al. 1999, Speier et al. 1999). Scientists who win awards read more on average (Tenopir and King).

Graduate students, particularly Ph.D. students, are often found to be heavy users of electronic journals (Rudner, Miller-Whitehead, and Gellman 2002; King and Montgomery 2002), most likely in their role as researchers. As found in the SuperJournal projects, graduate students may be “binge” users, consulting electronic journals extensively for a short period when they are writing a thesis or dissertation. There are some exceptions-undergraduates were the most frequent users in an experimental study conducted by the American Chemical Society (Entlich et al. 1996).

Different workplaces or types of institutions have varying use patterns as well. Davis (2002), as mentioned in Tier 1, examined electronic journal user logs in libraries of the NorthEast Research Library Consortium and found that each institution has a unique pattern of use-medical institution users had higher use of a smaller number of journals, while users at large universities and smaller colleges downloaded articles from a greater variety of journal titles.

Other differences are more controversial or less conclusive. Some studies (Speier et al. 1999; Tomney and Burton 1998; Antoir 2001; Monopoli et al. 2002; Palmer and Sandler 2003; Hahn et al. 1999) found differences in preferences or behavior based on the age of the user. In the late 1990s, Speier et al. (1999) and Hahn et al. (1999) found that younger ARL university business faculty members reported that they read from electronic journals more often and were more aware of electronic journals than were older faculty. In a study done in 1996 and 1997, more than half of the faculty members under the age of 40 reported using electronic journals, as compared with only 14% of those over 40, although more than 80% of the total respondents indicated they would consider using electronic journals in the future (Tomney and Burton 1998). Antoir (2001) found that older people preferred print articles; Monopoli et al. (2002) found that users between 21 and 34 used electronic journals most frequently. Older University of Michigan social sciences faculty members tended to prefer print more often than did younger faculty (Palmer and Sandler 2003). Age made a difference in how faculty members, staff, and students at Colorado State University rated their computer skills, with more respondents under 30 rating their skills as good (Cochenour and Moothart 2003).

Differences in electronic journal use may be attributed to age, status, or rank (Sathe, Grady, and Giuse 2002; Tenner and Yang 1999). Tenner and Yang (1999) found that assistant professors are most likely to have used electronic journals (44.7%), followed by full professors (34.5%), and associate professors (34.2%). Sathe, Grady, and Giuse (2002) found that undergraduates, medical students, and residents prefer electronic journals, while clinical and research faculty members prefer print. Researchers in other studies found no relationship between age and searching skills, although the researchers observed that younger users are more likely to browse on the computer, while older users prefer print journals for browsing (Brockman et al. 2001). Tenopir and King found no relationship between age and reading patterns among astronomers.

Age differences should be studied more, at least for the next generation, since most college students now are computer literate and report that they use the Web frequently (Tenopir 2003; Waldman 2003). This does not necessarily correlate with effective use, however. An experimental study by Cockrell and Jayne (2002) found that faculty performed tasks requiring retrieval and use of journal and newspaper articles significantly better than undergraduates. Tenopir (2003) found that students had more confidence in their searching skills than did their faculty. Although lower division undergraduate students used the Web and Web search engines frequently, students are mostly unaware of the distinctions between material on the Web and peer-reviewed journals. Freshman psychology students with higher self-efficacy scores are more motivated to learn about electronic journals and find electronic journals easier to use, but they also visit the library more often (Waldman 2003).

Gender differences are even more inconclusive. Only a few studies have examined gender as a factor in information use, beyond studies of recreational Internet use. Majid and Abazova (1999) found that male academic staff at the International Islamic University, Malaysia, reported they possessed better computing skills than did their female counterparts; Finholt and Brooks (1999) found males to be slightly more frequent users of JSTOR; and Monopoli et al. (2002) found that male staff at the University of Patras, Greece, reported more frequent use of electronic journals than did female staff (although a very small percent of students or faculty used electronic journals at all since the electronic journals are in English rather than Greek). These reported differences by gender are too minor, outdated, or idiosyncratic to form conclusive findings.

DLF/CLIR/Outsell (Tier 1) did, however, find differences between men’s and women’s reported use of online resources. More women than men say they use e-journals in their research; they reported using their institution’s Web site more often than men, and using electronic sources more than men most or all of the time. In teaching, men report significantly more use of a search engine to access e-journals than do women who access e-journals.

Some individual differences may be a factor only in early adoption and, as electronic resources become familiar and ubiquitous through the library, these differences may cease to be important. Other factors, such as the way different disciplines do their work, may be more pervasive. Mahe, Andrys, and Chartron (2000), for example, found that biologists rely on print because of deeply ingrained work habits, and earth scientists fail to see work advantages of electronic journals. Part of the hesitancy of some may be a discomfort with technology, unavailability of technology, or insufficient knowledge of electronic journals, all of which are already changing (Mahe, Andrys, and Chartron 2000; Curtis, Weller, and Hurd 1997). Physicists embrace preprints and have adopted e-print servers enthusiastically in part because their research tends to be highly collaborative and is conducted in large research institutions that have internal peer review.

3.2 Information-Seeking Behavior and Preferences

Students, faculty, and non-university professionals now use a variety of sources for articles, including electronic journals, print journals, Web sites of professional organizations, author’s Web sites, e-mail from colleagues, and e-print servers (Dillon and Hahn 2002; Tenopir and King; Cochenour and Moothart 2003). Print remains important for at least some information for all subject disciplines and as part of the research process for undergraduates (Dilevko and Gottlieb 2002). Even as early as 1996, academic users expressed dissatisfaction with library collections of printed journals, books, and conference proceedings (Bancroft et al. 1998).

Both current and older materials remain important, as many of the Tier 1 studies found. At the University of California, Berkeley, 93% of faculty members and 87% of graduate students across academic disciplines reported that they use materials older than five years “sometimes” or “often.” Most preferred electronic resources and often use the library from their desktops (Maughan 1999). Faculty and other subject experts make a distinction between core journal titles and non-core journals. At the University of Maryland, 70% of the faculty want core journals in both print and electronic form, but the same number wanted non-core journals only in electronic form (Dillon and Hahn 2002).

Although both browsing and searching remain important information-seeking strategies, electronic journals (in particular, full-text databases) are causing a decrease in browsing titles, while searching by topic has increased (Sathe, Grady, and Giuse 2002; Tenopir and King). Browsing of core journals by tables of contents remains important, but searching by topic for additional journals and articles is increasingly popular, particularly in large, mixed-journal title databases. Most libraries offer a combination of these large full-text databases, which facilitate searching, and journal systems from publishers, which facilitate browsing.

In systems restricted to journals from a single publisher, browsing through tables of contents remains important. Use of the John Wiley & Sons online journals in a biotechnology company was heavily weighted toward finding articles through browsing (1.42 average per session) as opposed to searching (.02 articles per session.) There was a dramatic increase in the number of articles viewed or downloaded between 1999 and 2001, but PDF remained a more popular format than HTML (Crawford 2002).

Worlock (2002) found that articles recommended by colleagues were more often in print than in electronic format. Contradictory results were obtained by Tenopir and King, who discovered that e-mail and listservs make it easier to share recommended articles with colleagues. University faculty members reported the ability to send articles to their colleagues instantly as being one of the major advantages of electronic journals (Palmer and Sandler 2003).

Use of online indexes and abstracts seems to have increased, particularly when there are links to full texts. More than 80% of veterinary medicine students at Iowa State University used online indexes in 1997, compared with only 16% a decade earlier (Pelzer, Wiese, and Leysen 1998), indicating a major shift to electronic resources during that decade. More than half of the medical faculty, residents, and students at a regional site of the University of Illinois report that they search MEDLINE at least weekly and much of their identification of journals comes from MEDLINE. Awareness and use of other abstracting and indexing services or full text databases is low, however (De Groote and Dorsch 2003).

The Pew studies (see Tier 1 analysis) and Graham (2003) clearly show that high school students, in particular, and undergraduate students prefer to search the Internet first for school-related tasks. When given a specific research task, only 2% of undergraduates at Wellesley College’s “Computers and the Internet” class included non-Internet resources in their answers. The students have extraordinary faith in their favorite search engine, even though they are unclear how it works.

A nationwide survey of students and academics in the Netherlands found that 60% of respondents in the humanities, 78% of respondents in the social sciences, and 82% of respondents in the sciences used the Internet for study or work and nearly all believed they had Internet skills (Voorbij 1999). Still, when rating the importance of different means of searching, more than 88% believed that subject searching of tables of contents databases was important or very important, followed by searching the OPAC, citations, and asking colleagues (asking a librarian came in last). Nearly two-thirds believed searching the Internet or the Web were important or very important, and most perceived that their Web searches yielded enough or more than enough information (Voorbij 1999). A vast majority reported they were self-taught Internet searchers, who rely on trial and error. Assistance from colleagues was the second most frequently cited means of acquiring searching skills; gaining the skills through library courses was far less commonly cited.

Ninety-seven percent of freshmen psychology students reported that they access the Internet at least weekly, about 44% of the time for educational information. More than three-quarters say they begin their research through the Internet, and two-thirds say they find most of their information through the Internet. Still, two-thirds also report that they visit the physical library at least weekly, most often for studying. The role of the library as a place to study and socialize (and sleep!) was more important than as a place to get information. Half of the students access the library’s electronic resources from home, and only a quarter said that use of the electronic resources was a reason to visit the library (Waldman 2003).

Dilevko and Gottlieb (2002) report efforts to attract undergraduates to the physical library. Since most undergraduates report that they turn first to online sources, and since turnstile counts were going down at the University of Toronto library, they surveyed students to find what role print materials still play for undergraduates. About 47% of undergraduates reported that they began their assignments with online sources 90% of the time, but printed journals and, especially, print books remain important in their research, particularly for humanities students. The authors conclude that print books are still vital and are associated with high-quality work. They recommend that librarians stress the value of printed materials in addition to online materials for the successful completion of assignments.

A teacher’s or librarian’s recommendation of specific sources, such as a library full-text database or a specific Web site, is reported to influence a student’s choice of sources (Waldman 2003; Tenopir 2003). At the University of California, Berkeley, the groups of faculty and graduate students that report the heaviest use of electronic resources also recommend the need for more library reference and instructional services (Maughan 1999). Use of electronic books is also clearly course driven-books with the highest usage are those required in a class, and most users come to an electronic book collection to use a single title (Summerfield and Mandel 1999).

MacDonald and Dunkelberger (2000) also found that most undergraduates in a composition class were always or almost always likely to use a recommended full-text database provided by the library as a first stop for information; the next largest group went first to the Web, and the smallest group went first to print. When searching on a mixed full-text/bibliographic database, many said they always restricted their search to full text. Most students did not limit their research to a single source of information, but of those who did, the Web was the most popular single source for information, followed by a full-text database (MacDonald and Dunkelberger 2000).

Faculty members and librarians can influence students’ choices of electronic resources, although faculty may not always be familiar with the range of sources available to them through the library. Library instruction, with time for practice, increases college students’ efficacy in online searching. Ren (2000) found that students who had a more positive attitude toward learning electronic information search skills had fewer negative emotions about electronic searching and performed better in assignments. Librarians should stress the importance of all library materials, including print and electronic resources, since undergraduate students value recommendations (Dilevko and Gottlieb 2002). The most effective way for students to learn about important resources in academic libraries seems to be for librarians to work directly with faculty to bring relevant electronic resources into the classroom (Tenopir 2003).

Students bring Web searching habits to their use of electronic scholarly materials and seem to have difficulty adapting to different types of information resources, interfaces, or search systems (Cockrell and Jayne 2002; Tenopir 2003). In a controlled study of 49 undergraduates, graduates, and faculty members, Cockrell and Jayne (2002) found that few undergraduates took time to read explanations or help screens, and that they give up easily and are not selective-they tend to choose the first item on the list, rather than scrolling down to see information displayed lower on the screen. These findings are similar to the many Web use studies beyond the scope of this report that show Web users in general tend to enter only a single search term and seldom look beyond the first screen (see, for example, Spink et al. 2001).

Students claim to use evaluation methods for Web sites (e.g., “If it’s pink with flowers it probably isn’t any good”) (Tenopir 2003). However, Graham (2003) found that students were susceptible to advertising claims on Web sites, government misinformation, and propaganda and could not consistently differentiate between advertising and fact. Only a few students in this experiment double-checked the information they found on the Web.

There is some evidence also that college students have a low tolerance for system features that don’t work or are too difficult. Bishop (1999) conducted user tests at the University of Illinois and found that if an abstract was missing when a student clicked on the abstract button, the student never again clicked on the button for abstracts. She concluded that one small system failure might have a long-term impact on student searching behavior.

Some college students report that they receive at least some training in evaluating library sources (57.6% in a study by Burton and Chadwick 2000), but a sizable minority do not. Those with no training show a slight preference for the Internet over the library for research. More than half of the undergraduate and graduate students from 97 different majors in a medium-sized Western university reported they use both the library and the Internet for research, while 21% use the library exclusively, 21% use the Internet exclusively, and 6% use neither (Burton and Chadwick 2000).

Although this report does not attempt to cover the many studies of Internet-only use and Internet searching behavior, these studies can provide some insights into the Web search patterns of children and adults who use the Web for both recreational and work-related purposes. To locate many of these studies, see Wang (1999), Yahoo’s directory under “Internet,” and Molyneux and Williams (1999).

3.3 Perceived Advantages of Electronic Resources and Preferences

Users perceive electronic resources-in particular electronic journals and, for students, the Internet-to hold many advantages. Faculty members at ARL institutions cited convenience, timeliness, and the ability to search text as the most important factors in choosing electronic journals over print (Lenares 1999). Least important to them was animation of graphics, although others sometimes mention that as an important advantage. In other surveys, graduate students said the top reasons for using electronic journals were the ability to link to additional information, the ability to search, and the currency of materials (Liew, Foo, and Chennupati 2000; Woodward et al. 1997). The ability to search across a wide range of journal articles, search within an article, and interact with multiple levels of information objects were listed as the top three significant features sought in future electronic journals (Liew, Foo, and Chennupati 2000).

Many studies have found that users believe the main advantage of electronic journals is convenience of accessing articles any time from their desktop computer (Palmer and Sandler 2003; Woodward et al. 1997; Rusch-Feja and Siebeky 1999; Maughan 1999; Tenner and Yang 1999; Hiller 2002; Nicolaides 2001; Chu 1998; Bishop 1999). Experienced users also liked the ease of skimming and searching, the possibility of downloading or printing the desired document or segment, the currency of information, the speed of access, and the ability to send articles to their colleagues instantly (Palmer and Sandler 2003; Rusch-Feja and Siebeky 1999; Sathe, Grady, and Giuse 2002; Entlich et al. 1996; Chu 1998). Storing articles electronically, then printing out a portable print copy, appeals to frequent e-journal users (Palmer and Sandler 2003).

Convenience and speed of access are mentioned or implied repeatedly. Students reported the top three ways that access to electronic resources has improved their academic careers: access to a wider range of information, faster access to information, and easier access to information (Ray and Day 1998). In England, Tilburg University faculty members cite timely availability, easy access, full text searching, and access from home as factors that promote the use of electronic journals (Roes 1999). Focus groups of engineering faculty members and students wanted to search electronic journals quickly and easily, but they desired interfaces that could be customized and the ability to create personal collections (Bishop 1995), while economics students and faculty want the addition of data sets (Nicolaides 2001).

The TULIP project, an early electronic journals study (1992-1995), was a cooperative undertaking between several university libraries and Elsevier. It attempted to predict the potential use of electronic journals through log analysis, focus groups, and interviews, while making sample collections available to faculty members and graduate students. The lessons from TULIP are incorporated into later commercial products, and the conclusions agree with later studies. Even in the early 1990s, faculty members and graduate students wanted electronic journal systems that are as intuitive as possible, preferably using a familiar interface, with access to all information from one source. They wanted high processing speed for downloading and printing, timely information, good image quality, many journal titles and sufficient dates covered, and linking. Graduate students used the system more often than faculty members. At this early date, TULIP researchers noticed an emotional tie to paper and the library-something that has diminished fairly rapidly with today’s convenience of electronic journals. Still, users liked the convenience of desktop access, but, consistent with almost all studies today, they preferred to print out a hard copy for reading. Promotion and training were both found to be crucial to develop a base of regular users (TULIP 1996). Familiarity, in the case of electronic journals, has bred continued use.

3.4 Problems or Concerns with Electronic Resources

Although the advantages are outweighing the perceived problems or concerns as use increases and more resources are available, users still express some concerns about the disadvantages of electronic library collections. Participants in several studies expressed the desire for more online materials, including additional journal titles, a wider variety of special or out-of-the mainstream materials, and complete volumes or back files of existing journals (Quigley et al. 2002; Palmer and Sandler 2003; Rusch-Feja and Siebeky 1999; Stewart 1996; Nicolaides 2001). Print is a proven archival format. Even those who prefer electronic access to journals (75% of respondents from the University of Michigan faculty in economics, sociology, and anthropology), prefer that books remain in print format (Palmer and Sandler 2003).

The most common complaint found in many studies is the discomfort of reading from the screen or poor graphic quality (Nelson 2001; Palmer and Sandler 2003; Woodward et al. 1997; Woodward et al. 1998; Sathe, Grady, and Giuse 2002; Costa 2000). Respondents consistently report that they prefer to print out articles for reading and do most of their reading from the paper printout (Stewart 1996; Entlich et al. 1996; Tomney and Burton 1998; Brown 1999; Woodward et al. 1997; King and Montgomery 2002; Cherry and Duff 2002; Duff and Cherry 2000). They prefer PDF format for printing, although the HTML format is better for skimming.

Faculty members from ARL institutions said that the most important characteristics that would lead them to choose print over electronic were ability to browse, portability, physical comfort, and convenience (Lenares 1999). In citing the chief reasons for preferring print over electronic journals, Vanderbilt University medical faculty and students said that print is an easier to read format, of better graphic quality, easier to browse, and easier to access (Sathe, Grady, and Giuse 2002). Access to adequate technology may still be a problem for some (Mahe, Andrys, and Chartron 2000); 22% of science faculty respondents at the University of Michigan requested that procedural or technological barriers to access be removed (Quigley et al. 2002).

When asked to identify problems, only a small percentage of respondents to most interviews or surveys agree to the same ones (or even agree that there are problems). The response rate for any one concern or problem is rarely more than 20% and “top problems” are usually expressed by less than 10% of the respondents. Students at one university were asked how “access to electronic resources has hindered your academic career.” Although not nearly as many agreed to hindrances as they did to improvements, the top three hindrances mentioned were that online access is time consuming (16.4%), it detracts from doing work (13.5%), and lack of information technology knowledge hinders effective use (11.1%) (Ray and Day 1998). The top category for disadvantages among faculty members and graduate students at Ohio State University was “don’t know” (24% and 33% respectively); an additional 8% of faculty and 6% of graduate students saw no disadvantages (Rogers 2001).

A continuing problem is that users may be unaware of relevant resources in the library collection. In a survey at the University of Maryland, 31% of the faculty members reported never using electronic journals; the reasons cited were unfamiliarity with how to access the journals and a lack of need because of personal subscriptions (Dillon and Hahn 2002). In a study of computer engineering undergraduate students in Nanyang Technological University in Singapore, researchers discovered that more than one-third of the respondents had never accessed computer engineering databases available through the library and of those, half had never heard of them (Majid and Tan 2002). Lower-division undergraduates in focus groups at the University of Tennessee report that they know the Web and major search engines such as Google, but unless a library resource is specifically named (and required) in a class, they are unaware of its usefulness (Tenopir 2003). Although French research scientists are using electronic journals more often, librarians still need to promote the resources because scientists hesitate to use electronic sources when they feel they have insufficient knowledge of them (Mahe, Andrys, and Chartron 2000).

The perception that electronic journals are of lower quality than print is another problem that may be diminishing as a high percentage of peer-reviewed journals are digitized. In the late 1990s, business school faculty members surveyed at ARL institutions reported that they did not perceive electronic journals to be of as high quality as paper counterparts; their responses changed, however, when they were asked to evaluate a well-respected print journal evolving to electronic format (Speier et al. 1999). While more than 70% of the faculty members in a British university believe the quality of articles in electronic journals is the same as in print journals, this same group of respondents cited the top disadvantage of electronic journals as being the impression that electronic publication is not “real” publication (Tomney and Burton 1998).

On the other hand, faculty members at the University of West England reported that they believed electronic journal content generally to be of good quality, and in some cases, to have added value. They said that they would use more electronic journals as more were made available in their area of interest and would recommend them to students (Nelson 2001). Texas A&M faculty members also reported that they have no objections to students using peer-reviewed electronic journals and that they would recommend electronic journals to students (Tenner and Yang 1999).

The proliferation of sources for articles and the sheer amount of information now available may be confusing to some users. Retrieving too much information is a problem mentioned by some, as is getting lost on a tangent and not knowing when to quit searching (Epic 2001). The distinction between the “article” and the “journal” in full-text databases was unclear to faculty members and undergraduates surveyed in the Decomate study (Nicolaides 2001), although at Columbia University (Epic 2001) researchers found that students clearly understood the difference between electronic databases and Web sites. In focus groups at the University of Tennessee, Tenopir (2003) found that students understand that information found on the Web is different from the resources provided by the library, but many are not fully aware of what resources the library offers.

3.5 Library Policies and Financial Concerns

Decisions that libraries make based on financial concerns, such as pay per use, may have unintended consequences on user behavior. Electronic journal collections in libraries are growing steadily (see, for example, Kidd 2002 and ARL 2003) and some libraries are formulating collection development policies that encourage lease of electronic journals over purchase of print journals (Montgomery and King 2002; King and Montgomery 2002; and Council of Australian University Librarians 2002). Even as early as 1999, 29% of ARL and 33.5% of non-ARL academic libraries reported cancellations of print journals in favor of electronic access, and more said they would cancel print in the future (Shemberg and Grossman 1999).

There has been a steady increase in the percentage of acquisitions dollars spent on electronic resources in ARL libraries. On average, ARL libraries spent 13.2% of their acquisitions budget on electronic resources in 1999Ð2000, and several libraries reported spending more than 20%. By 2000Ð2001 the average had grown to 16% (ARL 2003), and some special libraries spend a far greater percentage. Investment banking and brokerage firms are reported to spend 40-100% of their information budgets on online products, by far the highest of any type of company (Carrick 2002).

Library policies that favor electronic journals over print are having an effect on user behavior. Users are increasingly positive about electronic collections and visits to the physical library by faculty and graduate students are down in many libraries, replaced by visits to the virtual library (Rogers 2001; Hiller 2002). At the University of Washington, between 1998 and 2001 graduate students and faculty in the health sciences, sciences, and engineering reported the most pronounced decline in visits to the physical library. The primary use of the library by undergraduates tends to be as a workplace, although science and engineering students say they visited the library most often to find journals (Hiller 2002).

Libraries that report a decline in visits to the physical library as a decrease in library usage do themselves a disservice. Users enjoy the convenience and other benefits of electronic access and are adjusting their behavior as encouraged by library collection development policies. Total library use-physical plus virtual-is likely actually up in most institutions. Virtual library users are less likely to ask for help or communicate with librarians (Epic 2001), unless the library offers special virtual reference services.

University faculty members report that an increase in their electronic journal usage is accompanied by a decrease in the frequency of their use of print journals (Lenares 1999; Rogers 2001). Surveys of Ohio State University users from 1998 to 2000 found a steady increase in acceptance of electronic journals and their reported use. By 2000, almost two-thirds of faculty members and graduate students said it was important for OSU libraries to replace their print subscriptions with electronic journal subscriptions when permanent electronic storage is available (Rogers 2001).

Users of the University of Southern California Norris Medical Library viewed approximately 28,000 electronic full-text articles in a six-month period, as compared with 1,800 uses of the corresponding print volumes (Morse and Clintworth 2001). Even considering that users might have read more than one article per print volume, the electronic viewings far outnumbered the print. Part of this may be explained by the availability of new titles in electronic form, but use of both print and electronic titles was concentrated on a small number of the most popular titles-just 20 of the titles accounted for 60% of the total usage and the top 25 titles were common to both print and electronic (Morse and Clintworth 2001). The most requested electronic titles at the Elektronishche Zeitschriftenbibliothek in Germany are major journals whose print editions are also heavily used (Hutzler and Schupfner 2002). Use is also higher for electronic versions of books at Columbia University compared with the same titles available in print (Summerfield and Mandel 1999), although the University of Pittsburgh Health Sciences library found that use of both print journal titles and electronic journal titles covered by the Ovid online system increased at a similar rate (Tannery, Silverman, and Epstein 2002).

This pattern of a small number of titles accounting for a large percentage of use is, of course, not new to electronic resources. The so-called 80/20 rule has been well documented in library collections, where a large percentage of use is concentrated in a small percentage of the collection. Usage logs make the calculations of use much easier with electronic resources, and this 80/20 phenomenon is reported by many studies that use transaction log analysis (Roes 1999; Day 2001; Davis 2002). Interestingly, this rule may also hold true for users-a small percent of total library users is responsible for most electronic journal use (Entlich et al. 1996). This is similar to the phenomenon of “binge” users in the SuperJournal project. Davis and Solla found that a vast majority of users of American Chemical Society journals at Cornell University download few articles and consult few journals. They conclude that a small number of heavy users can have a great effect on the number of total downloads.

Still, as the OhioLINK studies have shown, many of the remaining 80 or so percent of the journals in an electronic collection will get some usage. In addition, Day (2001) found that at least one article was downloaded from 92% of the journals available to users at the University of Manchester Institute of Science and Technology. In his examination of NERL usage logs, however, Davis (2002) found that no institution uses every available title and some journal titles are used infrequently by all institutions. Overall, 90% of the downloads came from 40% of the collection.

Faculty and graduate students say 24-hour availability is a prime advantage of electronic journals (Rogers 2001), but even with 24/7 availability of the virtual library, most academic use follows the normal rhythms of the workweek and academic calendar. Just as turnstile counts mark the use of physical academic libraries, log data of virtual collections show peak use in March, November, and April (Mackie-Mason et al. 1999) and Monday through Thursday mid-morning to late afternoons, with a huge drop-off on Friday afternoons (Tenopir and Read 2000).

Although a user’s institution pays for subscriptions or access to electronic journals, this cost is hidden from the user. Any overt charge or obvious pay per-view has an impact on user behavior. The “Pricing Electronic Access to Knowledge” (PEAK) project in the late 1990s was a major experiment with 12 libraries of varying size and type and the Elsevier journal collection. It measured not only use of electronic journals by journal title and type of library, but also measured use under two different payment models for articles. Users of the subject libraries were provided with both “unmetered” access (in which access comes with subscription) and “metered” access (in which users receive an I.D. and use tokens, generally paid by the library, to get to full texts) to journal articles. Although use increased from the first to the second year in the experiment, 60% of accesses were for “unmetered” content, most of which was more than one year old. The study concluded that the “user cost of access, consisting of both monetary payments and time or effort, has a significant effect on the number of articles that readers access” (Mackie-Mason et al. 1999).

Pay per-view or pay per-use creates a barrier that affects the frequency of online access and downloads. Nicolas and Huntington (2002) found that users who entered an online journals system from a subscribing institution visited the collection more often than non-subscribers (who could search for free, but had to pay per article selected). Subscribers also spent more time viewing each article, viewed articles from more journals, and used a wider variety of journal titles and subjects than did non-subscribers. Although seven students studied at Central Connecticut State University found the end-user system Questia easy to use, only one of the seven thought “it is worth it to subscribe” on their own (Tomaiuolo 2001).

Some users may be willing to pay for electronic articles, at least part of the time. Worlock (2002) surveyed 252 working scientists and social scientists in the United Kingdom (split between academic and non-academic workplaces) to find out if they ever pay for articles. Nearly two-thirds of the respondents reported that, on average, they pay for between one and five articles per week beyond regular subscriptions. Still, two-thirds said they felt the articles were too expensive (Worlock 2002).

Passwords can be another barrier to use, in particular different passwords for different databases or collections (Roes 1999). Users want free (to them) access, without having to remember multiple passwords or log-on protocols. In eliminating special access requirements, however, libraries may create a problem for themselves. There is evidence that many faculty members and students do not realize that the numerous electronic journals they can reach from their office, dormitory, or home computer through their university user name or identification are actually paid for and provided by the library. If users are not aware what the library provides, they will be less inclined to advocate for the library at budget time.

Users can help libraries set collection and service priorities, with no one best solution for all types of libraries, types of users, or all subject disciplines. They can help make decisions about licensing choices and be made familiar with relative costs and tradeoffs. When asked to make choices about what they were willing to forego to get more electronic access, researchers from the Max Planck Society felt there were certain services or materials they could do without, including binding of journals, journals with low impact factors, and print versions of journals readily available electronically (Rusch-Feja and Siebeky 1999). Science, engineering, and health sciences faculty at the University of Washington favored canceling print journals in favor of electronic only, while humanities and social sciences faculty opposed this idea and responded that maintaining the quality of the print collection is their highest priority (Hiller 2002).

3.6 Summary of Tier 2

Many of the findings in the Tier 2 studies support findings in the larger Tier 1 studies. These consistent findings will help librarians know more about their users, which will help them set policy, make decisions, and design more effective products and services.

First, there is no one typical user and, thus, no satisfactory single information policy. Although all groups of users rely on electronic resources to some degree and will do so more in the future, the enthusiasm with which changes are embraced, the system features valued, and the need for continued print collections varies. These variations occur in different subject disciplines, but also in the way information is used, the task undertaken, and the role of the user. Currently, younger people may be more enthusiastic adopters of technology.

As libraries make more electronic journals and full-text databases available to users, both browsing and searching remain important information-seeking behaviors, but browsing by journal titles is decreasing while searching by topic is increasing. Most subject experts have a core group of journal titles that they browse, read from, and recommend to students, but they read from a wider variety of journals through subject searching. Most users employ a variety of sources to find journal articles, but high school and lower division college students most frequently turn first to the free Web and Web search engines such as Google or Yahoo.

Almost all types of users perceive many advantages of electronic journals, in particular when electronic journals are convenient for their work. The speed of access, desktop availability, and convenience of downloading and printing are most often mentioned as advantages.

Most users also perceive some disadvantages. Almost everyone prefers to print out articles in PDF format for reading, but to use HTML for viewing. Some novice users, or users where there is poor technological infrastructure, are worried about how well they will be able to use the technology. Some professionals are concerned about longevity and archiving.

Finally, library policies affect users in both anticipated and unanticipated ways. Visits to the physical library by faculty members and graduate students in particular decrease as more digital resources are accessible from their offices or homes. Undergraduates use the library as a place to socialize and study, so their frequency of use is affected less. When many additional journal titles are provided online, users will read from a wider variety of sources and read more, but most reading will still be done from a relatively small proportion of sources. This varies with subject discipline—medical users, for example, seem to read more from a smaller core group of titles. Barriers to use, including fees or passwords, will restrict use by almost all users except the most highly motivated.