cohen.p65 �������������������������������������275 A Com arison of Researgh Universit and T o Year Colle e Librar Web Sites Contenty F ngtionalit y and Form Laura B. Cohen and Julie M. Still The authors examined the library home pages of fifty Ph.D.-granting institutions and fifty two-year colleges to determine the home pages’ purpose as manifested by their content and its placement within the structure of the site. By looking at the library Web sites of contrasting academic institutions, the authors identified those aspects of the sites that reflect institutional character. They further identified a core common content that exists across academic library Web sites independent of the parent institution. eb sites are becoming a com­ monplace phenomenon in modern society, and academic libraries have been very much a part of this trend. Academic institutions were early contributors to the Internet ex­ periment, and it has long been a part of campus culture to provide computers that offer access to the resources contained in this global network. It is no surprise that libraries within academic institutions were early developers of Gopher and then Web sites. In recent years, library Web sites over­ all have undergone a tremendous prolif­ eration. As of October 1998, Libweb, one of the premier Web­based collections of links to library sites, listed nearly 2,500 Web sites in more than seventy countries l Despite the ubiquity of Web sites, both within academic institutions and more re­ cently in the social and business world, sys­ tematic research that studies the content, structure, and functionalities of these sites is only recently emerging. Three studies from 1996 illustrate the scope of this type of research, especially with respect to librar­ ies. Laurel A. Clyde compared school and public library Web pages in thirteen coun­ tries and identified a number of common content attributes among these sites.2 She stated that "There are many different rea­ sons or purposes for which a library might create a home page or Web site and to a certain extent these purposes determine the format and content of the page, al­ though most library pages or sites in the studies reported in this article appear to have been created to serve multiple pur­ poses. When this was the case, unless the pages/sites were well organized to lead Laura B, Cohen is Network Services Librarian at the University at Albany, SUNY; e-mail:LCOHEN@cnsvax,albany,edu, Julie M, Still is Multimedia Instruction Librarian at Rutgers Uni-versity, Camden; e-mail: still@crab,rutgers,edu, 275 276 College & Research Libraries people to the appropriate information or resources, the sites or pages could appear confusing and even chaotic." Mark Stove and Steven D. Zink compared the design of the main home page of forty higher edu­ cation library sites and echoed Olyde's re­ mark about the poor design of some of the sites encountered. They also identified common elements to the main landing page, including the title, name of sponsor­ ing entity, description of purpose, index­ ing links to content within the site, and con­ tact information.3 Jan Whalen studied a small number of college and university li­ brary Web sites to assess content. The pages studied featured links to library services, resources on the Web, directions for locat­ ing the library, details of access for the dis­ abled, and a few special features such as lists of library instruction classes and a vir­ tual tour of the library.4 In 1998, David L. King investigated the main, front-end home page of all 120 ARL It may be assumed that any librarymounting a Web site does so withinthe context of an institution thatoffers Web access on publicworkstations or has expectations thata critical mass of its users hasInternet access from off-campus. library Web sites and identified common elements of the "typical" ARL library home page in the areas of page background, document headers, document footer, graphics, hypertext links, unlinked text, and page length.5 Among his findings was the frequent appearance of identifying titles, the tag to the author/cre­ ator, text­based hyperlinks to home page content, and text consisting of welcome messages and/or descriptions of hyperlinks The average main screen con­ tained approximately twenty­two hypertext links A useful article on Web page organiza­ tion was written by Steven W Sowards, who identified depth criteria to describe the design of ready reference sites of mostly academic library home pages 6 Group I sites May 1999 place all their links on a single page. Group II sites have two layers in which a click on the primary screen connects to one of sev­ eral secondary pages, which contains links connecting to the resources. Group III sites contain three or more layers of pages in a more complex hierarchy. Sowards further identified subtypes within these three groups. An academic library functions within the context of its parent institution. This has great implications for what can and can­ not be accomplished by a library wishing to sponsor a Web site. The mounting and maintaining of a professional Web site of any kind can require a considerable alloca­ tion of resources. This can include invest­ ment in hardware, software and network access for one's constituency, end­user training and support, staff time and exper­ tise for content development and mainte­ nance, and development and maintenance of the infrastructure that hosts the site.� It may be assumed that any library mounting a Web site does so within the context of an institution that offers Web access on public workstations or has expec­ tations that a critical mass of its users has Internet access from off­campus. It also may be assumed that such libraries create sites in the expectation that they will pro­ vide information helpful to, and wanted by, their constituents. Given this context, the investigators be­ gan this study with the knowledge that the nature and size of academic institutions are reflected in the nature and size of their li­ braries. This, in turn, should be manifested somehow in their libraries' Web sites. The purpose of the current study was to inves­ tigate the academic library Web sites of con­ trasting institutions in order to analyze any similarities and differences between them in three major areas: the purposes these Web sites serve as manifested by their con­ tent; the functionalities offered; and the structure of the sites. A study of this type has not been done. The authors chose to compare the Web sites of Ph.D.­granting research universi­ ties and two­year colleges because these may be seen as two ends of the institutional spectrum. An investigation of this type would lend itself to the identification of those aspects of a library's Web site that reflect institutional character. Further, such a study in contrasts would be helpful in identifying a core common content that may exist across academic library Web sites, independent of the parent institution. A brief profile of one school of each type investigated in this study illustrates the contrast. For example, the University of California at San Diego enrolled 18,119 stu­ dents in 1997 and includes seven separate academic units. The twelve libraries on campus contain a total of 2.3 million books and 23,906 periodical subscriptions.8 By comparison, Niagara County Community College enrolled 5,118 students in 1997. The school has one library containing 85,767 books and 538 periodical subscriptions.9 Although vastly different in size and scope, both types of institutions are impor­ tant to higher education. In 1995, there were 3,706 higher education institutions in the United States, 1,462 of which were two­ year schools and 115 were Ph.D.­granting institutions.10 Enrolllent in Ph.D.-granting institutions totaled 2,999,641, whereas 5,492,529 students attended two -year schools. Roughly half of first-year fresh- men (1,017,970 out of 2,168,831) attended a two -year school� About one quarter (26.38%) of those attending college in 1995 went to a two-year school. About one third of African-American students (32%), and more than a third of Hispanic students (38%) were enrolled in two-year schools.I3 There is an overlap in the usage of these institutions in that students may attend a two­year school and later transfer to a larger institution. This overlap suggests another important reason to compare the Web sites of these two types of institutions. Transfer students often manage to elude library instruction at the larger schools because most general library instruction takes place during fresh­ man year or in conjunction with basic courses. The information provided on li­ brary Web sites may be the only type of li­ brary orientation that transfer students re­ ceive. Previous studies of library services Comparing College Library Web Sites 277 to transfer students were conducted before the advent of Web sites.14 Methodology To gather the data, one author selected fifty Phgag-granting institutions and the other fifty two-year collegesg To cover the great- est possible geographic diversity, the origi- nal intention was to choose one library of each type from each state; however, this proved problematicg Therefore, some states are represented by more than one type of library and other states are not represented at allg However, the representation is equal for each state for each type of libraryg A full list of Web sites examined is available on the WebgIs The authors created a worksheet to col­ lect data for each library Web site. They communicated with each other through­ out the data collection process, making ad­ ditional comments on the sheet as seren­ dipitous items of interest were found. The items reviewed can be grouped by cat­ egory: content, functionality, and structure. Some of the larger, more complex research library sites had links to a dozen or more libraries and affiliated research centers. The investigator took cursory looks at a mini­ mum of two individual libraries' pages but followed through all levels of only one, the most easily identifiable general library (as opposed to smaller subject libraries). Data collection took place during July 1998. Content In addressing the purpose of these library sites as identifiable by their content, the authors asked the following four ques­ tionso 1. Is the library Web site an information tool? A Web site can provide a library with an opportunity to disseminate information about itself. This can include items such as physical address, staff directory, descrip­ tion of services, departments and policies, and announcements of library news and events. A Web site also might provide infor­ mation about itself. This may take the form of an update date, a contact e­mail link, and a mission statement or disclaimer. 2. Is the library Web site a reference tool? http:sites.14 http:schools.I3 http:institutions.10 278 College & Research Libraries The World Wide Web contains a wealth of reference material. One indication of this is the presence of links to reference sources such as biographical information, dictionar­ ies, and subscription encyclopedias. Links to Web­based search engines such as AltaVista and Infoseek were counted, as were links to Web directories such as Ya­ hoo!, The Internet Public Library, the WWW Virtual Library, etc. Some libraries also use their Web sites to provide infor­ mation on the local and regional area to their constituencies. Finally, links to a library's OPAC and the OPACs of other li­ braries were included in this category. 3. Is the library Web site a research tool? A library Web site can contribute to the library's effort to address the research needs of its academic community. Included in this category are subject bibliographies of Internet resources in academic disci- plines (as opposed to general reference). It was assumed that these subject pages were compiled by the library's staff, and noted when a page resided at another institution. Also in this category were pathfinders to locally held resources. These pathfinders may exist first as hard-copy library hand- outs before being mounted on the Web. With the proliferation of free periodicals on the Web, links to these resources also were identified and counted. Finally, the inves- tigators noted the presence of subscription or paid Web-based services other than en- cyclopedias, such as the indexes and full-text resources provided by FirstSearch, IAC Searchoank, JSTOR, Lexis-Nexis, and Project Muse. 4. Is the library Web site an instructional tool? A library Web site is an attractive me­ dium for the dissemination of instructional materials. The investigators looked at Internet/Internet software tutorials, tuto­ rials on the use of the library, and computer software tutorials. Library use tutorials might include instructions on how to use an index, a guide to database search strate­ gies, how to write a bibliography, how to locate articles, elements of the research pro­ cess, and so on, but not help screens within a given tool. The presence of software tu­ torials might indicate the blurring of the May 1999 lines between an academic library and aca­ demic computing. For most of the items described above, the investigators made a count of the num­ ber of links included for each-for example, the total number of reference sources linked from a library's Web site. It was sometimes a challenge to identify these sources. On some of the Web sites, refer­ ence sources were linked from the same jump­off page as academic subject areas, or mixed in with other types of sources on the same page. Every attempt was made to identify and count resources in any lo­ cation where they were found. There also were cases in which the same link was made from multiple pages. As the amount of duplication appeared euite small, there was no attempt to avoid a double count. Functionalities To address functionalities, the authors looked at two technologies commonly found on Web sites: the use of a site search engine, and interactive request forms. A site search engine allows the user to enter a search statement that the engine then will process against the files on that site and re­ turn a list of hits. The ability to conduct such a search can offer users the opportunity to locate materials of interest without need­ ing to navigate through the site. This is es­ pecially useful on large and complex sites such as those provided by some of the re­ search libraries. Interactive request forms allow users to submit various types of re­ quests by filling out a template and sub­ mitting the form. Examples for library Web sites include requests for interlibrary loan (ILL) and the purchase of library material. Structure A Web site has a structure, or architecture, that consists of a main landing screen, or page, and sfghyui90­pages that are ac­ cessed via hyperlinks from the main page. Beyond the main page, a Web site can in­ clude different types of information struc­ tures.16 Because of the hypertext nature of the Web, this structure can be quite com­ plex. A particular page or function may be approachable from multiple locations within the same site. For example, an OPAC may be linked from several pages within a library's site. The authors recorded the top­ most appearance of each item investigated (i.e., the level closest to the main page). This would reveal the amount of navigation away from the main page that was neces­ sary to access a particular item. In many cases, multiple access points throughout the site also were noted. Site architecture across six levels were examined, starting with the main screen: • Main Screen: An item appeared on the landing page. • Level 1: An item was linked from the landing page. • Level 2: An item was linked from a sub­ sequent page; and so on to level S. At this distance from the main page, very few resources made a first appearance. In fact, two­year college Web sites rarely reached this level of com­ plexity, whereas on some research library sites users do not begin to access the actual content of branch library pages until reach­ ing level 3, and new content could appear several levels below that. Results Restust Cvery item investigated appeared on the Web sites studied, and all but one appeared on the sites of both types of schools. Table 1 shows the total number of library Web sites that contained the resources studied. A fuller picture was achieved by looking at the total number of occurrences of individual items in the reference, research, and instruction categories. To assist in comparisons, the to­ tals for selected resource types were orga­ nized into clusters (e.g., 1-S items, 6-10 items, and so on). These totals illustrated the degree to which the items were covered. These data are presented in tables 2 and 3. Taken together, these tables provided an­ swers to the four questions about the pur­ poses these academic library Web sites serve. The study's first question asked whether the Web site served as an information tool. Most of the Web sites studied provided some degree of information about their li­ braries. The most prevalent information of­ Comparing College Library Web Sites 279 fered was a contact e­mail link, a Web page update date, and a description of services. More than half of the research libraries included all the items except a Web page statement and separate description of poli­ cies. Measuring a separate description of services, departments, and policies turned out to be unrealistic for the research librar­ ies because these descriptions usually ap­ peared together in various combinations. The actual coverage for policies, therefore, was higher than the separate number in­ dicates. It was striking to note that only five(10%) libraries of each type includeda mission statement or disclaimerabout their Web sites. braries included a contact e­mail link, up­ date date, staff directory, description of policies, and description of services. Given the small size of these organizations, it was no surprise that only two of these libraries offered departmental descriptions. More­ over, there was very little coverage of li­ brary news and events (nine, or 18%). Al­ though this may be surprising from a pub­ lic relations point of view, it might make sense when considering the staff time needed to keep such an item current. Of the thirty­five two­year college library sites that provided an update date, six had last been updated ten or more months prior to the investigator 's visit. This also may be a factor in the relative sparsity of news and events information. It was striking to note that only five (10%) libraries of each type included a mis­ sion statement or disclaimer about their Web sites. Thus, these libraries allow the public to make assumptions about the mis­ sion of their sites, which may be contrary to the libraries' intent, and also overlook the opportunity to publicly absolve them­ selves of any liability as to their sites' con­ tent or the content linked from their sites. This also may be an indication that, as many case studies have shown, it is not uncom­ mon for library Web sites to be a "bottom­ up" project, begun because someone on the More than half the two­year college li­ staff had the knowledge to do so, not be- college sites linked to extensive meta-sites 280 College & Research Libraries May 1999 TABLE 1 Total Web Sites Containing Resources Resource Ph.D. Percentage Two-Year Percentage Library information Web page statement 5 10 5 10 Contact email link 50 100 39 78 Physical address 30 60 24 48 Update date 44 88 35 70 Staff directory 36 72 26 52 News/Events 44 88 9 18 Description of departments 32 64 2 4 Description of services 40 80 44 88 Description of policies 15 30 36 72 Reference Links to search engines 38 76 32 64 Links to Net directories 22 44 23 46 Reference section 39 78 37 74 Links to own OPAC 50 100 42 84 Links to other OPACs 44 88 33 66 Links to local info 23 46 25 50 Subscription encyclopedia 41 82 14 28 Research Links to Net subject resources 42 84 30 60 Pathfinders to local resources 30 60 5 10 Links to subscription services 49 98 36 72 Links to free periodicals 23 46 19 38 Instruction Net/Net software tutorials 10 20 26 52 Library use tutorials 21 42 14 28 Software tutorials 0 0 3 6 Functionalities Request forms 35 70 15 30 Search engine 29 58 3 6 cause library administration (or anyone else) developed a mission or guidelines for the site. Such a statement may come later or not at all. The second question asked was whether the library Web site served as a reference tool. Again, all the items studied appeared on the Web sites of both types of institutions. Most prevalent were links to a library's own OPAC. All the research libraries and 84 per- cent of the two-year college libraries offered these links. There also was a high occurrence of links to the OPACs of other libraries, with 88 percent of the research libraries and 66 percent of the two-year college libraries making these links. Two of these two-year of OPACs compiled at remote libraries, whereas two others linked to meta-sites cre- ated by consortia of which they were mem- bers. Most two-year college library sites linked to ten or fewer other OPACs. Many of the research libraries also linked to meta-sites such as the Libweb site men- tioned earlier. Another common theme was a link or links to the OPAC of other libraries in the geographic area, although not neces- sarily as part of a consortium. Almost an equal number of these librar- ies, thirty-nine research libraries and thirty- seven two-year college libraries, either of- fered a separate section of reference sources or included reference links on their sites. The degree of coverage, however, tells a more complete story. Overall, the two­ year college libraries did a minimal job of providing access to reference sources on the Web. As table 3 demonstrates, fifteen of the thirty­seven libraries that included reference sources linked to twenty or fewer of them, whereas only eight linked to more than 100. Five of these libraries relied pri­ marily on reference pages located on other libraries' Web sites. This practice can be Comparing College Library Web Sites 281 beneficial to libraries without the resources to produce an extensive listing on their own. An interesting example was the com­ pilation of reference links at the DelAWARE site, created as a joint project of the Dela­ ware Division of Libraries, Delaware Pub­ lic Libraries, and Delaware Technical & Community College. Both types of libraries were more likely to link to Internet search engines than to Internet directories. Seventy-six percent of TABLE 2 Clusters of Total Resources for Ph.D. Schools Resource Clusters of Total Links Reference 1–20 21–40 41–60 61–80 81–100 101–150 151–200 201–250 Links to 6 2 3 5 4 10 4 2 reference sites 1–5 6–10 11–15 16–20 21+ Links to search engines 5 23 5 4 1 Links to Net directories 16 5 0 1 0 Links to other OPACs 15 19 3 2 5 Links to local info 9 6 5 0 3 Research 1–5 6–10 11–15 16–20 21+ Links to Net subject resources 2 3 3 2 32 Pathfinders to local resources 2 1 2 1 24 Links to subscription services 0 1 1 3 44 Links to free periodicals 3 3 3 0 12 251-300 2 TABLE 3 Clusters of Total Resources for Two-Year Schools Resource Clusters of Total Links Reference 1–20 21–40 41–60 61–80 81–100 101–150 151–200 201–250 Links to 15 3 6 3 2 5 1 1 reference sites 1–5 6–10 11–15 16–20 21+ Links to search engines 6 15 5 3 3 Links to Net directories 14 4 3 2 0 Links to other OPACs 21 6 1 2 3 Links to local info 11 3 2 2 7 Research 1–5 6–10 11–15 16–20 21+ Links to Net subject resources 15 3 6 1 5 Pathfinders to local resources 2 2 1 0 0 Links to subscription services 27 5 3 0 1 Links to free periodicals 10 5 2 0 2 Instruction 1–5 6–10 11–15 16–20 21+ Net/Net software tutorials 16 5 0 2 3 Library use tutorials 10 3 1 0 0 Software tutorials 3 0 0 0 0 251-300 1 282 College & Research Libraries the research libraries made linns to search engines as did 64 percent of the twoyyear college librariess Linns to Internet directoy ries appeared on only 44 percent of the rey search library pages and 46 percent of the twoyyear college pagess It was notable that all of the libraries that linned to Internet directories pointed to twenty or fewer diy rectoriesk most in both categories linned to five or fewers A number of the Web sites of both types of libraries presented inaccurate lists of search engines by including the Yay hoo! directorys This was the case with eleven research libraries and nine twoyyear college librariess The popularity of Webybased subscripy tion encyclopedias was especially strining for the research libraries 82 percent of which had linns to this type of resources In contrast only 28 percent of the twoyyear college libraries linned to a subscription encyclopedias By far the most popular eny cyclopedia was Britannicas This subscripy tion was represented on thirtyyeight rey search library sites and eleven twoyyear college library sitess There also was a scaty tering of subscriptions for Cambridge Eny cyclopedia Encyclopedia Americana Acay demic American Encyclopedia arolier es and Encartas It was unusual for a site to linn to more than one subscription encyclopey dias Only four Web sites of research library ies and one twoyyear college site linned to more than one subscriptions About half these libraries used their Web sites to provide linns to local or regional iny formations Of those providing linns of this type linnss more than half offered fewer than ten Another aspect of a Web sitees effectivey ness as a reference tool is its offering of any notationss It is one thing to mane a linnk it is quite another to supply a descriptive or evaluative blurb about the linnedyto sites Such an annotation can provide a helpful guide to users as they scan lists of linns and choose those sites they wish to visits Table 4 demonstrates that annotations were in short supply on the pages investigateds For example only fourteen of the thirtyysix rey search libraries (3u%) linning to search eny gines annotated these linnss gine of thirtyy May 1999 two (28%) twoyyear college libraries annoy tated these linnss The situation was more extreme for Internet directoriess Four of twentyynine research libraries (14%) annoy tated these linns compared with six of twentyythree (26%) twoyyear college library iess These resources vary significantly from one another but users of these library Web sites are being left to discover the differy ences for themselves without the benefit of a librarianes assessments A similar situay tion held true for the linns to reference sourcess Six of fortyyone research libraries (1u%) linning to reference sources annoy tated these linns whereas twelve of thirtyy six (33%) twoyyear college libraries made annotationss Although a few libraries (more often research libraries) included a section on evaluating search engines in their library tutorials these screens usually were not linned to the search engines pages Linewise tutorial sections on evaluating Web matey rials in general usually are not accessible than the tutorial itself or even mentioned from any other point s The third question asned whether the library Web site served as a research tools As might be expected the Web sites of rey search institutions did well in this categorys Subscription database services often pury chased through consortia or on a systemwide basis were a major avenue for offering Webybased research materials to both types of librariess Table 1 shows that fortyynine research libraries provided linns to subscription database services fortyyfour of which linned to more than twentyyone of these toolss Among the twoyyear college libraries thirtyysix offered these services twentyyseven of these subscribing to five or fewer servicess Two of the libraries were members of OhioLinn and The Virtual Liy brary of Virginia (VIVA) consortia that ofy fered them access to twenty or more suby scriptionss The efforts of such large consory tia offer these schools a tremendous wealth of Internetybased research toolss It was iny teresting to note that a number of these sites acnnowledged the consortial arrangey ments Further twoyyear college library sites were more linely to linn to subscription sery vices by name (ises FirstSearch SearchBann Comparing College Library Web Sites 283 etc ) than research library sites which tended to link to the individual databases contained in these services sometimes within larger lists of free databases and journals arranged alphabetically by sub- ject or other combination At some sites especially the two-year college sites it is possible to determine what services or ven- dors the databases were purchased from by running the computer mouse over the link Many research libraries however have loaded the databases directly on their server so that the mouse displays a local link and not a link to the vendor Another popular avenue for providing research materials was the mounting of discipline-specific pages of links to Internet resources created by library staff Forty-two (84%) research libraries offered such pages thirty-two of which created pages in more than twenty subject areas Thirty (60%) of the two -year college libraries offered discipline-specific pages of links to Internet resources half of these providing pages in five or fewer subject areas Four of these libraries relied on subject pages created at other libraries Although not counting the number of links on these subject pages one investigator noted that pages created by li- brary staff in these two-year colleges usu- ally contained a very small number of links sometimes only a hand- ful for each subject area The size and complexity of these resources at re- search libraries varied greatly Although a number of subject areas were represented some of these sites linked to only a few individually chosen tools; others linked to multiple meta-sites compiled at other institutions This is not always evident from the resource page unless one checks the address of each link individually Given the enormity of this task for the pur- poses of this study only the number of links was counted and not necessarily the type of link (local or remote meta-site) Although many libraries included Internet subject lists on their sites there were variations on the use of annotations Fifteen of thirty two-year college libraries included annotations (table 4) A correct number was more difficult to establish for the research libraries and thus is not in- cluded in table 4 Because the Internet sub- ject resource lists may be on individual li- brary pages instead of subject pages or may be compiled by librarians from the various libraries there is very little consistency within institutions and sometimes within pages The listings on one subject page may be annotated but those on another may not be; or some individual items on a list are annotated and others are not Those librar- ies with a clear consistent policy would appear to have decided against annota- tions This is yet another example of the way the complex organization of the re- search university impacts on the libraries and therefore on Web page organization ,athfinders to locally held resources were less common appearing on 60 percent of the research library sites and only 10 percent of the sites for two-year colleges (table 1) Many of the research libraries that offered path- TABLE 4 Annotations Resource Total Annotated Percentage Ph.D. Links to search engines 36 14 39 Links to Net directories 29 4 14 Links to reference sites 41 6 15 Links to own OPAC 50 10 20 Links to other OPACs 43 10 23 Links to local info 24 2 8 Two-Year Links to search engines 32 9 28 Links to net directories 23 6 26 Links to reference sites 36 12 33 Links to own OPAC 42 1 2 Links to other OPACs 33 1 3 Links to local info 25 3 12 Links to Net subject resources 30 15 50 284 College & Research Libraries finders, however, did so as a significant ef­ fort. Twenty­four of the thirty libraries mounting this type of resource offered path­ finders in more than twenty research areas. By contrast, four of the five two­year col­ lege libraries offering pathfinders did so in ten or fewer areas. It was interesting to note that twenty­three research library sites of­ fered both subject resource pages and path­ finders, and only two offered neither. The overall sparsity of pathfinders was surpris­ ing, especially in the case of two­year col­ lege libraries. Most libraries offer printed pathfinders of some form or another. Given that many word processors now have the capability of saving text documents as HTML files, it would be relatively easy to produce at least a few pathfinders as Web documents. The lack of Web­enabled path­ finders, however, may be an indication of staffing level rather than ability. It is possible that at many libraries either no one has had the time to do this or the library administra­ tion has not made it a priority. Some of the research librariespresented separate listings forsubscription electronic journals andfree titles, and others organized themall in one long listing. Forty­six percent of the research librar­ ies and 38 percent of two­year college li­ braries took advantage of the free periodi­ cals on the Web by linking to them (table 1). Twelve of the research libraries linked to more than twenty of these periodicals, whereas ten of the two­year college librar­ ies linked to five or fewer. One notable ex­ ception was the two­year college library that linked to the Web site of every peri­ odical it owned that maintained a site; this amounted to more than 200 links. Some of the research libraries presented separate listings for subscription electronic journals and free titles, and others orga­ nized them all in one long listing. Because some Web­accessible OPACs now have the capability of including links within their catalog records, it is possible that some libraries have cataloged certain free electronic journals and included links there, May 1999 instead of compiling a separate listing (al­ though some libraries have done both). Finally, this study investigated whether the library Web sites served as an instruc­ tional tool. The picture here was some­ what mixed. The two­year college librar­ ies offered Internet or Internet software tutorials more frequently than research li­ braries, 52 percent as compared with 20 percent. Given the greater division of du­ ties at larger schools, it is possible that In­ ternet­related tutorials are available, but on the computer center 's site instead of the library's site. Of the ten research li­ braries that offered these tutorials, half linked to tutorials created by other librar­ ies. Nineteen of twenty­six two­year col­ lege libraries did the same. Even so, six­ teen of these libraries linked to five or fewer of these tutorials. Library use tuto­ rials were mounted by 42 percent of the research libraries and 28 percent of the two­year college libraries (table 1). Of the fourteen two­year college libraries offer­ ing these tutorials, ten mounted five or fewer. The preponderance of library use tutorials on the research library sites sug­ gests that large, complex libraries or li­ brary systems create the need for tutori­ als, have the staff to create these tutorials, and are oriented toward publishing tuto­ rials on the Web. It is useful to note that the availability of these tutorials can ben­ efit transfer students who miss out on for­ mal library orientations or instruction. If librarians are taking on the role of teaching computer software, they are not mounting these tutorials on the Web. None of the research libraries mounted software tutorials, compared with only three two­ year college libraries. Once again, it is pos­ sible that these tutorials are available else­ where in the institution's Web structure. Functionalities It was no surprise that a greater number of research library sites offered Web-based functionalities than did two-year college li- brary sites. It may be assumed that research libraries have access to larger academic com- puting units with more resources to exploit and share, or they may have the technol- ogy and staff to implement these functionalities themselves. Interactive re­ quest forms were present on the sites of 70 percent of these libraries, the most common type being ILL. Fifteen sites (30%) of the two­year college libraries offered interactive request forms. Found in almost equal num­ bers were forms for ILL, requesting an item for purchase, asking a reference question, and submitting a comment to library staff. A search engine was offered on 58 per­ cent of the research library sites. As some of these Web sites were of substantial size, the search engine was an appropriate of­ fering. The smaller sites of the two­year col­ lege libraries had little need of a search en­ gine. This was reflected in the fact that only three of these sites offered one. Site Structure It was immediately apparent that the type of school had a dramatic impact on the size and arrangement of the Web site. Two­year college library sites encompassed the main page through three levels down, with the second and third levels introducing little new material. Only one resource type-free electronic journals-extended as far as four levels away from the main screen on just one library site, and these were linked from subject resource pages that were organized from levels 2 and 3. It was notable that some of the two­year college sites were extremely small. Four of them consisted of only one page, one of two pages, and three of three pages; these ac­ counted for 16 percent of the two­year col­ lege sites studied. On these limited sites, the main screen served an extremely im­ portant function. For example, one library site placed all thirteen links to reference sources on the main screen, and another did the same with its eleven reference links. Two of the other smaller sites also placed all of their reference links on the main page. This is in contrast to the more usual prac­ tice on larger sites of placing one link to ref­ erence sources on the main screen, with the links to individual sites appearing at the next one or two levels down. These links were recorded as level 1 in both situations because they were linked from the main screen, but Comparing College Library Web Sites 285 it is useful to note that four small two­year college sites used their main screen as a catch­all for most, if not all, of their resources. Certainly the main screen of any Web site serves a vital function. When looking at the research and two­year college library sites together, it can be seen that this land­ ing page serves two major functions: (1) as an index with links to information located on subsequent pages as represented by level 1 in this study, and (2) as a container of information. Its function for level 1 link­ ages has been discussed earlier in this sec­ tion. Looking at the main page as a con­ tainer, the researchers found that the two­ year college Web sites made a somewhat greater use of this page to present infor­ mation. This may be seen as a reflection of the smaller, less hierarchical nature of these sites. Table 5 shows the comparison of the use of the main screen for both types of li­ braries. Only items categorized as library information were found here. It is notable, for example, that three two­year college sites described their policies here, six de­ scribed their services, and four placed their staff directory on the main page. None of the research library sites used the main page for these purposes. Use of the main page for placement of the update date and physical address was common for both li­ brary types. These latter findings are in general agreement with those of King in his study of ARL library sites. The two­year college library sites, there­ fore, were organized primarily around the main screen and then one level down as linked from this page. Links from the main screen, as represented by level 1, accounted for a significant presentation of the content of these two­year college library sites. All of the content relating to reference, re­ search, and instruction appeared on level 1 on at least 65 percent of the library pages offering this content. This is illustrated in table 6. Some two­year college libraries created sites for limited purposes. The most bare­ bones site consisted of a description of li­ brary services. Another contained a de­ scription of library services, a staff direc­ tory, and a link to an EBSCO database. Yet 286 College & Research Libraries another contained a description of services, links to thirteen reference sites, a link to its own OPAC and to two other OPACs, and links to two local sites. These pages served the functions of information, reference, and research, but only in the most minimal way. Seven other libraries were similarly limited in scope. Research library Web sites, on the other hand, tended to be very complex and often were reflective of the complexity of the in­ stitution. Some sites were easy to navigate, but others were labyrinthine in nature and difficult to maneuver without getting lost. Most of the institutions studied had more than one library, and thus the main screen primarily included materials that were ap­ plicable systemwide along with links to in­ dividual library pages. These individual li­ brary pages also included links back to the main screen or to items included on it (such as the OPAC). Thus, there was initial dupli­ cation and cross­linkage. Update dates at some institutions were not given on the main page, but only on secondary and sub­ sequent pages. Pathfinders and staff direc­ tories were often on the in­ dividual library pages, not on the system main screen. Sometimes pathfinders or guides to Internet re­ Resource Total Total Main Percentage sources were located on the Ph.D. Reference Department Web page statement home page of an individual Contact email link library, thus appearing in Physical address Update datethis study as level 3, while Staff directory it would be located at level News/events2 to someone starting at that library 's page and Description of departments 32 0 0 Description of serviceslevel 1 for someone who Description of policieshad bookmarked the de­ partmental page. More­ Two-Year over, there was clearly a Web page statement lack of systemwide Web Contact email linkpolicies, as some libraries at Physical addressthe same institution in­ Update datecluded items (such as path­ Staff directory finders) that others did not. News/eventsFurther, there was little May 1999 larger Web sites of research libraries were more evenly distributed through level 3 and thus were more hierarchical. Items ap­ pearing most significantly on level 1 (i.e., linked from the main screen) were a con­ tact e­mail link, the library's OPAC, request forms, and the site search engine. Table 7 shows the items that were linked from the main screen by more than half the research libraries offering these items. It was notable that all of the research li­ brary sites presented instructional materi­ als starting at level 2 and beyond. For ex­ ample, four of ten libraries offering Internet tutorials presented them on level 2, whereas nine of the 21 sites offering library use tutorials placed them on this level. This is related to the arrangement of these sites using the main screen as a systemwide screen. It was rare to find instructional material at this level; it was most often linked off of an individual library's page, perhaps reflecting a lack of systemwide agreement on the use or importance of gen­ eral library instruction. Another factor may be that librarians develop instructional ma­ TABLE 5 Use of Main Screen 5 1 20 50 0 0 30 16 53 44 25 57 36 0 0 44 5 11 40 0 0 15 0 0 5 3 60 39 0 0 24 11 46 35 34 97 26 4 15 9 0 0 Description of departments 2 0 0uniformity of arrange­ Description of services 44 6 14ment. Description of policies 36 3 8 Items on the relatively Comparing College Library Web Sites 287 TABLE 6 Links from Main Screen (Level 1) by Two-Year Schools Resource Total Total Level 1 Percentage Library Information Web page statement 5 2 40 Contact email link 39 36 92 Physical address 24 7 29 Update date 35 0 0 Staff directory 26 19 73 News/events 9 9 100 Description of departments 2 1 50 Description of services 44 29 66 Description of policies 36 22 61 Reference Links to search engines 32 24 75 Links to Net directories 23 15 65 Reference section 37 30 81 Links to own OPAC 42 39 93 Links to other OPACs 33 27 82 Links to local info 25 17 68 Subscription encyclopedia 14 13 93 Research Links to Net subject resources 30 22 73 Pathfinders to local resources 5 4 80 Links to subscription services 36 33 92 Links to free periodicals 19 13 68 Instruction Net/Net software tutorials 26 18 69 Library use tutorials 14 11 79 Software tutorials 3 2 67 Functionalities Request forms 15 10 67 Search engines 3 2 67 terials on a local level, and what is useful at of new content that appeared on level S is one library may not be applicable else­ where within the system or campus. After level 3, the first appearance of content significantly dropped off. Thir­ teen items made a first appearance on level 4, whereas only seven made a first appearance on level S. On level 4, the item appearing most significantly was the col­ lection of links to Internet directories, ap­ pearing here on 38 percent of the sites that offered this content. This is a fair distance from the main page, although if one fol­ lows the path from systemwide page to individual library page to departmental page to finding tools page to Internet di­ rectories, it does make some sense, al­ though perhaps not to users. The sparsity illustrated in table 8. Discussion and Conclusions Dne masor goal in examining these fifty re­ search library and fifty two­year college li­ brary leb sites was to discover the pur­ pose these leb sites serve as manifested by their content. These sites might serve four possible purposese as a tool for infor­ mation, reference, research, and instruc­ tion. Considered as a group, the libraries studied served all these purposes, although each individual site may not have. Indi­ vidual research library sites, however, were more likely to fulfill all of these purposes, and to fulfill them to a far greater degree, than the sites of two­year college libraries. 288 College & Research Libraries Research library sites contained a greater quantity of what they offered than two­year college libraries. For example, in the case of subject­related Internet re­ sources, forty­one research library sites cov­ ered more than twenty subject areas, whereas half the two­year college libraries offering this type of resource covered five or fewer subject areas. Two­year college li­ brary sites also were less likely to offer in­ structional materials on library use, include pathfinders for topical research, or mount their own Internet tutorials. Coverage of reference sources was sparse. Certain in­ dividual two­year college library sites were so small that they fulfilled a very limited purpose in a very limited way. Whether these limitations were intentional or neces­ sitated by a lack of resources is a question that must be asked and one that could have been answered had these sites included a mission statement. The lack of such a state­ ment is especially compelling as half these sites either had no update date or had not been updated in about a year. A small number of two­year college li­ brary sites did create their own reference, research, and instructional materials. Over­ all, however, two­year college sites offered substantial resources only when they linked to the materials and meta­collections mounted by other, larger library sites. Given the hyperlinking nature of the Web, in which site creators are encouraged to link to external resources, it was surprising that this practice was not more widely en­ countered. Such a practice also can offer access to materials that a small library can­ not collect. For example, it may be accepted that students at these colleges have as May 1999 much need for Internet reference sources and tutorials on Internet use as their coun­ terparts at research institutions. High­qual­ ity collections of these resources are not dif­ ficult to find or, if need be, to adapt to local conditions. In the course of this research, the au­ thors searched for a core content that was common to these academic library Web sites. This core was defined as the presence of the item in more than half of the library sites studied. Table 1 demonstrates that the following items constitute a core: contact e­mail link; update date; staff directory; description of services; links to search engines; reference section/links; links to the library's own OPAC; links to other libraries' OPAC; links to Internet subject resources; links to subscription database/journal services. Considering the research libraries sepa­ rately, the following items can be added: physical address; library news/events; description of departments; links to a subscription encyclopedia; pathfinders to local resources; request forms. Considering the two­year college librar­ ies separately, the following items can be added: description of policies; Internet/Internet software tutorials. This research has shown that there is a core common content that can be identi­ fied for academic library Web sites indepen­ TABLE 7 Links from Main Screen (Level 1) by More Than Half of Research Libraries Resource Total Total Level 1 Percentage Contact email link 50 44 88 News/events 44 32 73 Description of services 40 22 55 Links to own OPAC 50 43 86 Links to subscription services 49 26 53 Request forms 35 19 54 Search engine 29 24 83 Comparing College Library Web Sites 289 TABLE 8 New Content Appearing on Level 5 Resource Total Ph.D. Total Level 5 Percentage Links to search engines 38 1 3 Links to Net directories 22 1 5 Reference section 41 1 2 Links to local info 24 1 4 Subscription encyclopedia 41 1 2 Links to Net subject resources 42 2 5 Pathfinders to local resources 30 3 10 Links to subscription services 49 1 2 Net/Net software tutorials 10 1 10 dent of the parent institution when re­ itself. This was manifested in the research search universities and two­year colleges are compared. The general purpose of an academic library Web site is to serve as a tool for information, reference, and re­ search. The departure occurs when one considers scope and scale. The two­year college sites offer fewer resources within a smaller, simpler architectural scheme. The complexity of a Web site, which may re­ quire the presence of a site search engine, is related to the complexity of the library and, on a broader scale, of the institution library sites. These larger institutions, for example, make greater use of freely avail­ able Internet resources and are more likely to create Web versions of paper documents (pathfinders). As a result, patrons of the more complex libraries-and their Web sites-are asked to select from and use a larger number and variety of materials and to navigate their way through a larger maze of resources. That these Web sites offer a greater number of tutorials is highly appro­ priate. Notes 1. Thomas Dowling, Libweb. Available at .2. Laurel A. Clyde, "The Library as Information Provider: The Home Page," Electronic Library 14.6 (Dec. 1996): 549-58.3. Mark Stove and Steven D. Zink, "World Wide Web Home Page Design: Patterns and Anoma­lies of Higher Education Library Home Pages," 4. Jan Whalen/ AA Study of Library Web Sites" (Nov. 1996). Available at Reference Services Review 24 (fall 1996): 7-20..5. David L. King, "Library Home Page Design: A Comparison of Page Layout for Front-ends toARL Library Web Sites," College & Research Libraries 59 (Sept. 1998): 458-65.6. Steven W. Sowards: 'A Topology of Ready Reference Web Sites in Libraries:' First Monday 3(May 1998). Available at f7. Peter A. Bloniarz and Kai R. Larsen, 'A Cost Performance _odel for Assessing WWW ServiceInvestments" (June1997). Available at . 8. Peterson's Register of Higher Edtcation 1998, 11th ed. (Prinueton, N..: Peterson's, 1998): 687. 9. Peterson's Guide to Two-Year Couueges 1998, 28th ed. (Princeton, N. J.: Peterson's,1998 ): 464. 10. Digest of Education Statistics 199l, NCES 98-015 by Thomas D. Snyder et al (Washington, D.C. U.S. Department of Education, National Center for Education Statistics, 1997) 259.11. Ibid., 183.12. Ibid., 191.13. Statistical Abstract of the United States: 1996, 116lh ed. (Washington, D.C.: U.S. Bureau of the Census, 1996): 190.14. Jennifer Cox and Ralph Johnson, "Transfer Students in the Library: The Forgotten Popula­tion," Research Strategies 10.2 (spring 1992): 88-91.15. The list of research university libraries is located at .The list of two-year college libraries is located at . 16. Patrick J. Lynch: "Site Structure:' in Web Style Manual (Yale Center for Advanced Instruc­tional Media, 1996). Available at .17. The DelAWARE site is available at . http://www.ctg.albany.edM/projects/inettb/SpreadSheets.html http:7-20.