Previous   Contents   Next
Issues in Science and Technology Librarianship
Winter 2001

URLs in this document have been updated. Links enclosed in {curly brackets} have been changed. If a replacement link was located, the new URL was added and the link is active; if a new site could not be identified, the broken link was removed.

[Refereed
article]

Web Sites of Science-Engineering Libraries: An Analysis of Content and Design

Nestor L. Osorio
Northern Illinois University
DeKalb, IL 60115
nosorio@niu.edu

Abstract

The purpose of this project is to identify trends in the design and content of home pages for web sites of Science-Engineering Libraries. Forty-five web sites from universities in the USA and Canada are chosen and their design characteristics and hypertext links analyzed. In addition, full-text journals, Internet resources and access to e-journals through the online catalog are examined.

Using the program Excel, data collected from each home page are tabulated. The most common elements are then incorporated into a model home page that visually shows the predominant design characteristics and hypertext links on this type of web site.

The results show that in general the current design of home pages for Science- Engineering Libraries contain many of the elements found in home pages of academic libraries. Among the characteristics found are images, screen lengths, colors, number and types of links, and link headings. The content of these web sites is also analyzed and summarized.

Introduction

In the last ten years, since the introduction of the World Wide Web as a vehicle for making information available to people, academic libraries have taken the lead by establishing a strong presence on the WWW. The web site has become one of the academic library's most visible artifacts for communicating with users, and the home page has become the starting point for users conducting library research to explore the resources available. Web sites of academic libraries vary in form and content depending on the nature of the university, its mission, size and the kind of programs offered. A sub-set of this very large community consists of the group of web sites dedicated to the support of science-engineering libraries. This study concentrates on this subset, and an analysis of its web sites is undertaken. The purpose of this study is to examine the home pages of science-engineering libraries, to identify their principal characteristics in terms of design features and content configuration and to determine current trends.

A technique that has been used in engineering for a long time is prototyping. With the support provided by computing technology, prototyping has now expanded to other fields, in particular, to information systems. Basically a prototype is a working model. According to Nickols (1993) it "provides a much closer picture of the system to be developed than an entire library full of user requirements, systems specifications, data dictionaries, functional flowcharts and memoranda...." In the area of library web development, use of this technique has not been widely reported. A keyword search in the Library Literature database from FirstSearch provides no results. This technique has been introduced here to give a visual representation of how the information collected is represented in a model form.

A brief review of the literature provides examples of the type of work being done in this area; the presentation of experimental design and methodology follows. Results of the study are then presented, and discussion of the results and conclusions complete the study.

Literature Review

As is stated by Cohen and Still (1999), a body of literature pertaining to the systematic study of web sites' content and structure is developing. In their study the authors compared the library home pages of research universities with those of two-year colleges in order to find contrasting differences in three areas: content, functionality and structure. King (1998) examined the home pages of libraries in the Association for Research Libraries (ARL) to determine web page front-end design. In one of the earlier studies of this nature, Stover and Zink (1996) created a ten-design elements matrix to review the home page of forty academic library web sites. An interesting article by Sowards (1998) discussed a theoretical typology of reference sites based on their depth (steps), organization and searching features. Within a more general scope, two recent articles have been published about the design of home pages; Dalal, Quible and Wyatt (2000) argued for the need to write cognitive guidelines for the design of better home pages based on an experimental study. And, after mentioning how little is known about web site design, Huizingh (2000) presented a framework consisting of two parts: content and design. Huizingh used this framework to investigate how different groups of corporations use the web. Arguments for further studies about how the design, structure and content of web sites can achieve best results is commonly found in the current literature on building web sites for corporate, organization or academic institutions (e.g., Abels et al. (1997), Misic and Johnson (1999), and Haas and Grams (2000)). For those interested in additional readings, D'Angelo and Little (1998), as well as Stover and Zink (1996), offer extensive literature reviews on how to build web sites. Finally, Sroka (1986) provides a detailed description of the technique of prototyping and its uses in developing information systems.

Methodology

Forty-five web sites of Science Engineering Libraries were selected for this study. These web sites were obtained from the Engineering Library Division (ELD) site. ELD is a division of the American Society for Engineering Education (ASEE). ELD's {web page} contains a list of 118 institutions, 75 of which are classified by the {Carnegie Foundation} as DR Ext or DR Int. There are 22 other U.S. institutions smaller in size, nine large academic research institutions from Canada, and the other 12 are from government agencies or from corporate libraries. The 45 selected include institutions from different geographical locations in the USA and Canada, private and public universities, top research universities, as well as institutions recognized because of their teaching emphasis. For the purpose of this project the list from the ELD page is considered comprehensive and the selection made is representative of the target group. An important selection criterion was that the libraries selected have a physical entity separate from the institution's main library. Of the selected libraries, 26 are science and engineering libraries and 19 are engineering libraries. Only one site per school was chosen; when a school has listed separate science and engineering libraries, the engineering library was chosen. All the data were collected during the month of April 2000.

The analysis of these web sites concentrates on the science-engineering libraries' home pages. After an initial exploration of each home page, two main areas of data collection were identified: 1. design characteristics of the home page; and 2. content on the home page as reflected by hyperlinks. This second section has three components: general hyperlinks, specialized hyperlinks, and hyperlinks to other sites in the university.

Internet research is similar to working with a moving target. In a content analysis of web sites carried out by Haas and Grams (2000), the sites to be analyzed were frozen in time by downloading their entire content using specialized software. In that way changes that occurred on the real site during the time of data gathering did not interfere with the experiment. In this study, the technique of downloading the content of the web site for their analysis was not used. Instead, all data were collected during a short, one-month period (April 2000). The sites were monitored to check for major overhauls. This technique worked well for data collection, but it was noticed that significant changes in two web sites had occurred after all the information about them was collected.

Some design characteristics include use of images, tables, background color, side bars, use of fonts, navigation bars, and hyperlinks. In King (1999), home page length, backgrounds (image or color), graphics, and hypertext links were some of the design features analyzed. Images, background, and color were also investigated by D'Angelo and Little (1998). In the general content area, examples of hyperlinks include hours, online catalog, services, and search engines. Examples of specialized links include patents and trademarks, standards, and products. Links to other university sites include the main library and the university web site.

Since the terminology used to designate headings of hyperlinks is not consistent, an effort is made to cover the different variations of titles used for the same kind of link under one generic heading. For example: the heading "Services" can be found as "User Services" or "Public Services". In this article, for the sake of simplicity, the term "topic heading" is referred to as "heading".

Each of the 45 home pages was analyzed; the data collected form a matrix of 66 elements identified in each home page. A record was kept for each library. The Excel spreadsheet was used to collect the data; a number one or a blank was assigned to designate whether the page had the feature or not. The total number of occurrences for each feature was calculated and a simple percentage was obtained. In four instances when the percentage was not meaningful, for example for "number of colors used", a simple average was calculated.

Prototyping models for the design of academic web sites have not been reported in the literature but some of the potential advantages such a model can have include the involvement of people in the design process and the appropriate feedback received in terms of needs and requirements. Identifying the essential elements of the system is the base for creating a mode (Sroka 1986). In this project, in order to create a model home page, every feature evaluated having a percentage of about thirty or higher was considered a predominant feature and is incorporated into the model or prototype.

The next part of this project was to investigate trends in providing links to full-text journals, the use of online catalogs for access to full-text journals, and links to Internet resources.

The entry hyperlink heading for full-text journals was identified, a record was kept of home page entry points, and the route of links was followed in order to find the electronic journals. Also tabulated were the ways these journals were presented, for example, by subject, in alphabetical order, by provider, etc. A similar analysis was done for Internet resources. For the OPACs, the path to the online catalog was recorded, and a title search of an e-journal held by the library was done. A record of the title was displayed and checked to see if the record indicated that this title was available in electronic format. The presence in the record of a hypertext link to the journal or provider was verified.

Results

According to Diaz (1998), a library web site can play different roles; it can serve as a workstation where a user finds databases, electronic texts, and the online catalog. It is a way to make library-made products available, it is used as a window to the WWW by making Internet resources available on a selective basis, and it is a communication tool where information about services, people and facilities and collections can be found. The results of this survey show that most science-engineering libraries are using their web sites to cover all these roles.

Table 1
Home Page Design
# of Occurrences Percentage
Links 45 100%
Fonts 42 93%
Navigation bar 30 67%
Library photograph/logo 24 53%
Side bars/table 24 53%
Other images 17 38%
Three port frameset 13 29%
Institutional logo 12 27%
Link select box 9 20%
Underuse of images 8 18%
Multi-frame tables 6 13%
Overuse of images 5 11%
Background images 3 7%
Background color 1 2%
Number of links avg. 34.08
Number of colors avg. 2.82
Number of screen lengths avg. 2.13
Number of images avg. 2.11
Links on Home Page (General)
# of Occurrences Percentage
Hours 35 78%
Online catalog 34 76%
About the library 33 73%
Electronic databases 30 67%
Where is the library (location/address) 27 60%
Search this site 23 51%
Library news 22 49%
Services 22 49%
Resources by subject 22 49%
Staff directory 21 47%
Page feedback, Comments, Suggestion box 20 44%
ILL forms 18 40%
Electronic resources 17 38%
Internet resources 17 38%
Collection description 15 33%
Use the library or tour, Library guides 14 31%
Ask a librarian 14 31%
Library reserves 14 31%
Find sci/tech information 12 27%
Other libraries 10 22%
Search engines 10 22%
Policies and procedures 9 20%
Faculty liaisons 9 20%
User education (program) 9 20%
New books 8 18%
FAQs 7 16%
Subject guides 6 13%
Recommend purchases 6 13%
The electronic reference desk 6 13%
Photocopies 6 13%
Searchlight (to search across databases) 5 11%
Help (to use the home page) 5 11%
Site index or map 5 11%
Sci/tech news 3 7%
CD-ROM workstations 2 4%
Senior design projects 1 2%
Specialized Links
# of Occurrences Percentage
Full-text journals 27 60%
Patents and trademarks 12 27%
Standards 8 18%
Map collection 7 16%
Journals guide 4 9%
Full-text books 4 9%
Audiovisuals 4 9%
Products 2 4%
Links to Other University Sites
# of Occurrences Percentage
Main library 34 76%
University web site 28 62%
College of Engineering/Science 19 42%
Distance education 3 7%

The top section of Table 1 provides data related to the design characteristics of the home pages. The first three features found -- links, fonts and navigation bar -- are almost considered standard in web pages, therefore, their high percentages (100%, 93% and 67%) were expected. The fact that about one half of the pages have a picture or a logo of the library is indicative of trying to give the page a visual identity with the library. With easy access to all kinds of graphic software, this feature is becoming a trend. Further, it is noted that the lower rate for inclusion of other images (38%) and institutional logo (27%) might be related to the fact that because the average page is 2 screen lengths, designers have to make selective choices about which images to include. The three port frameset feature was present in nearly one third of the pages. Its popularity derives from the fact that this feature allows for the distribution of hypertext links into compartments or levels. Users can identify links to the left side of the page; the text is concentrated on the main portion of the page, but there are links in that area as well. The number of colors (average 2.82) as a design element should be taken into consideration. Using a large number of colors might create visual confusion or distraction. This concept can also apply to the number of images (average 2.11). It appears that web designers were recognizing this important detail but still making their pages interesting.

There are differences present in the length of a single page. Pages with only one screen length do not have enough space to present essential links and information, but in the other extreme, multiple screen pages (four or more) seemed difficult to navigate and might cause users to lose interest. It appears that a home page with about two screen lengths is a reasonable solution. This result agrees with research previously done (Stover 1996; King 1998). The length of the page is directly related to how much text is included. For example, a brief description of a service or a link is acceptable but long explanations about, for example, a data base should be reserved to a special section in the instructional pages. King (1998) reported that the average number of links in a page was 22 with a range from 1 to 81. Stover (1996) has reported a page having 300 links. In this study the average number of links is 34. If all the links identified in this study as listed on Table 1 were included in a home page there would be 84 in total. A page with this number of links may be hard to navigate. Web design allows for distributing content into levels; the home page or front-end is the first level. A good design practice is to present key and essential links in level one and to distribute the content of the web site into further levels. Having 34 links in a home page of two screen lengths appears to be an acceptable design strategy.

The second section of Table 1 provides data about the most common general links available in a home page. Links present in over 65% of the web sites are Hours, Online Catalog, About the Library, and Electronic Databases. This is an indication that the main functions of the home page are to provide information about the library and immediate access to some of the most important databases. Surprisingly, 18 web sites do not tell users up-front where the building is located. A search engine for site searching is a feature that is becoming standard, 51%, and that number is expected to change drastically in the near future since it helps users to navigate a web site. In general, the data in this section can be divided into three types of links: links dealing with general information; research oriented links; and instructional links. A major concern is the fact that the lack of standard definitions makes analyzing links a difficult task. For example, ILL Forms (40%) can be included under Services or can appear under Use the Library/Tours/Guides. A similar case occurs for Library Reserves. Finding Information about Sci/Tech topics can be included with Resources by Subjects and vice versa. It is noticeable that about one third of the sites examined are using a form of electronic reference.

Section three of Table 1 provides data about the most common links to specialized resources present in a home page. In this case only two links were predominant: Full-text Journals and Patents and Trademarks. By examining the different levels of the web sites it is found that it is a matter of web organization. The difference in percentages between the two links (60% and 27%) derives from the fact that there are other avenues to find information about patents and trademarks. They are more likely to be found under other headings, such as Internet Resources, Resources by Subject or Find Sci/Tech Information. Specialized resources are usually found at a deeper level of the web site organization.

The most common links found related to other institutional sites are basically considered standard links: Main Library, University web site and links to a college or a department (76%, 62%, and 42%). An interesting question is why the first two are not 100 percent? The idea of having a site of a branch library without a link to the Main library is a rather peculiar one. Most likely these links are found in a secondary level of the web site. It can be said that links to these institutional web sites can always be found.

Further discussion of results not included in Table 1 follows. They are based on the examination of Internet resources, full-text journals, access of e-journals through the online catalog, levels of navigation, and databases listed on the home page.

Hypertext access headings for Internet Resources are numerous. Sixty-five headings as Internet resources links on home pages are found. Twenty-three libraries use more than one heading link as an entry point. Four libraries do not have links on their home page for Internet Resources. The number of entry headings varies from zero to seven. Twenty home pages have one entry heading, and eight home pages have two entry headings to access Internet resources. Often (23 cases), resources listed as Internet resources are a combination of free web sites and a library's databases. In some cases, the Internet Resources link provides users with tutorials or information about standards, patents and trademarks.

The results show that 45 hypertext (topic) headings are used on home pages as an entry point to access full-text journals. The most common headings are: Electronic Journals (13), and E-Journals (4). Most home pages require two clicks to reach a list of e-journals. Most home pages provide multiple access points to the e-journals. The number of entry headings varies from one to five. Twenty-eight home pages have one entry heading, and ten home pages have two entry headings. Alphabetical lists of full-text journal titles with hypertext links are common; other ways to present e-journals are by publishers, societies and providers or vendors.

Six libraries' home pages do not have direct links to a catalog or to a system's digital library. The OPAC record identified a title as an e-journal in 40 cases. The OPAC record had a hypertext link to the journal or provider in 34 cases. Only 14 home pages had a one-click (step) access to the search engine of the library OPAC. Finally, 15 home pages needed two clicks, and six home pages needed three clicks to reach the search engine of the OPAC.

Topologies of the web site structures in this study are complex. For both Internet resources and electronic journals, two clicks are usual and three clicks common. Access to the same information is usually obtained through multiple paths. For example, the heading "Electronic Journals" on a home page can branch out into four access points: by subject, category, collections (vendors) and in alphabetical order. It can get more complex when the user first has to go to the main library web site to access databases.

Ten libraries have electronic databases available directly from their home page. The list below does not include items appearing as new electronic products often shown in the News column. The databases appearing on the home page of some libraries are: Academic Search, ACM Digital Library, Agricola, Applied Science & Technology Index, Beilstein CrossFire, BIOSIS, CA Student Edition, Cambridge Scientific, CINAHL, CISTI Sources, Community of Science, Compendex Web, Composites Industries Abstracts, CRC Handbooks Online, Current Contents, Digital Dissertations, EBSCO Host, Ei Village, FirstSearch, GeoRef, Global Mobility Database, Health Ref Center, IEEE Explorer, IMechE Journals, INSPEC, Journal Citation Reports, Lexis-Nexis, Materials Properties, MathSciNet, Medline, Metadex, NCSTRL, NTIS, ProQuest, QUAKELINE, Sci Search, Science Citation Index, Science Direct, Science Magazine, Science's Next Wave, SciFinder Scholar, Soc Sci Search, TRIS Online, UnCoverWeb, and Web of Science.

[Sample home page]

Model home page

This model is based on the results presented in Table 1. It includes the design characteristics and the hyperlinks of contents that are predominant (30% or more) in the group of home pages studied and it serves as a visual representation of the characteristics found.

Discussion and Conclusions

This study has examined the design features and content elements of home pages at 45 science-engineering libraries. The results show that these home pages have many of the elements found in other academic home pages; it was also found that they have the problems and limitations typically found in similar home pages. It is not intended to be an evaluation instrument of web sites but rather a tool to create the fingerprints of typical science engineering web sites.

Unlike other studies where a critical tool is developed and a scale of values is defined in order to evaluate home pages, (e.g., in Stover and Zink (1996) and Huizingh (2000)), this study does not contain evaluation criteria nor scales to measure features. Neither, in this study are a number of elements identified, nor groups of libraries defined and then compared as in Cohen and Still (1999). This study is an exploratory tour to identify predominant design features and content elements. Sixty-six elements with at least one occurrence are found, 18 elements are related to the design of home pages, 48 elements are related to the content of the web sites. It is arbitrarily decided that if an element, for example, "navigation bar" or "staff directory" is present about 30% of the time or more then this element is considered a predominant element. The model page shows the predominant elements in today's home pages of science engineering-libraries based on the sample of pages selected.

Design Elements

Dewey (1999) indicated that one major problem in locating services and resources on a web site is the terminology currently used. Obscure labels (topic headings), labels that are too general or too specific are often found in many home pages. In this study, the use of multiple terms to indicate the same resource or service was noted on the web sites analyzed. It gets more confusing when, within a web site, different terms are used as headings to designate the same source. It was reported before that 49 terms are used to indicate access to electronic journals; some of the terms used such as "electronic journals" and "electronic journals databases" are precise. Other terms used like "electronic publications" and "electronic resources" can include more than one type of publication, and terms like "finding information" and "engineering research guides" can be obscure. A similar situation was noted for accessing Internet resources. There are also problems in using the terminology at different web sites; the term "electronic resources" in one site accesses indexes, e-journals, and newspapers. In another site it accesses only index databases, and on a third site it is used for linking to the OPAC and databases (indexes). A similar term "electronic databases" is used in one site to access indexes, in another site to access specialized full-text databases. In conclusion, an effort has to be made to standardize the terms used as topic headings.

The problem of multiple layers in web sites is one of the major concerns Dalal (2000) mentioned. The fact that users experience disorientation when they move into the deeper layers of the site structure is compounded by the large number of links to internal and external pages. For that reason, several authors recognized that more than three layers of text are considered excessive. In this study, two and three layers are very common, but there are instances of four, five, and even seven layers.

Content Elements

Cohen and Still (1999) have suggested that a web site is a tool for presenting four major library functions: information, reference, research and instruction. How do the results of this study fare against this concept? Data gathered in this study were organized with different groups of content elements -- general, specialized and links to other university sites. The data shown in Table 1 clearly indicate that content elements of the functions suggested by Cohen and Still were well represented. Also, in the model of predominant features, these four content functions were also present. Therefore, it is possible to conclude that today's home pages of science-engineering libraries are the gateways to information, reference, research, and instruction for the institutions they serve.

It has been said that content is the most important aspect of a home page. In this project, it is observed that in most cases the text included in the pages under examination is concise and clear, but contrary to previous research done (King 1998), no attempt is made to count the lines of content shown in each page. Another aspect of content includes the resources made available through the home page. Table 1 shows the effective use of library home pages to promote and provide access to valuable and scholarly information to the students and researchers they serve. In the general content aspect of these web sites, important information about the library such as hours open, a description of the library, library news, services offered, location and direction to the facilities, staff directory, collection description and ILL forms is made available. The research and educational aspect is found by having hypertext links, for example, to the online catalog, electronic databases, Internet resources, full-text journals, and patents and trademarks. The instructional purpose of the web site is revealed by directing users on how to find information about science and technical topics, having electronic resources arranged by subjects, and by providing library guides and tours. In addition, other content elements included are access to the main library, the university web site and to a college or department. In general terms, it is possible to conclude that science-engineering libraries are making an effort to make their home page a content-rich environment.

Further Research

Are the predominant design features and content elements found in this project the best features and elements to be incorporated into a science/engineering web site? It would be interesting, for example, to identify which home pages have the most characteristics represented in the model page, then design an evaluation instrument to measure the quality of these home pages as compared to home pages that are not like the model page. The results would be a benchmark (Misic and Johnson 1999) that would present the best practices of current home pages for science engineering libraries.

The author would like to recognize Peter Osorio, Marmion Academy, Illinois, for his work in building the model home page, and, Aaron Champene and Ted Larson from Northern Illinois University Libraries for their contributions in the spreadsheet design.

Bibliography

Abels, E.G., et al. 1997. Identifying user-based criteria for web pages. Internet Research 7 (4): 252-262.

Cohen, L.B. & Still, J.M. 1999. A comparison of research university and two-year college library web sites: content, functionality, and form. College & Research Libraries 60 (3): 275-289.

Dalal, N.P., et al. 2000. Cognitive design of home pages: an experimental study of comprehension on the World Wide Web. Information Processing & Management 36 (4): 607-621.

D'Angelo, J. & Little, S. K. 1998. Successful web pages: what are they and do they exist?. Information Technology and Libraries 17 (2): 71-81.

Dewey, B.I. 1999. In search of services: analyzing the findability of links on CIC university libraries' web pages. Information Technology and Libraries 18 (4): 210-213.

Diaz, K.R. 1998. The role of the library web site: a step beyond deli sandwiches. Reference & User Services Quarterly 38 (1): 41-43.

Haas, S.W. & Grams, E.S. 2000. Readers, authors, and page structure: a discussion of four questions arising from a content analysis of web pages. Journal of the American Society for Information Science 51 (2): 181-192.

Huizingh, E.K.R.E. 2000. The content and design of web sites: an empirical study. Information & Management 37 (3): 123-134.

King, D.L. 1998. Library home page design: a comparison of page layout for front-ends to ARL library web sites. College & Research Libraries 59 (5): 458-465.

Misic, M.M. & Johnson, K.L. 1999. Benchmarking: a tool for web site evaluation and improvement. Internet Research 9 (5): 383-392.

Nickols, F.W. 1993. Prototyping: systems development in record time. Journal of Systems Management 44 (9): 26-30.

Sowards, S.W. 1998. A topology of ready reference web sites in libraries. First Monday 3. [Online]. Available: {http://firstmonday.org/htbin/cgiwrap/bin/ojs/index.php/fm/article/view/594/515} [May, 1998].

Sroka, J.M. & Rader, M.H. 1986. Prototyping increases chance of systems acceptance. Data Management 24 (3): 12-20.

Stover, M. & Zink, S.D. 1996. World Wide Web home page design: patterns and anomalies of higher education library home pages. Reference Services Review 24 (3): 7-20.

Previous   Contents   Next

W3C 
4.0 Checked!