Web Exclusive

Coverage of ASIS 1997 Annual Meeting

Providing Library Resources via the World Wide Web: Implications of User Studies

ASIS Annual Meeting Technical Session sponsored by SIG LAN, November 5, 1997

Oya Y. Rieger, Cornell University Library: User Behavior in the electronic Library: The Case for Adaptive and Flexible Interfaces
Dan Iddings, Carnegie Library, Pittsburgh: Library Online Catalog Use in the Web Environment: The Experience of a Public Library
Laura Cousineau and John Little, Duke University: Where Our Patrons Go in Cyberspace: Click-Stream Analysis in an Academic Research Library
Cassandra Armstrong, Pittsburgh Theological Seminary
Gwyneth Duncan, Perkins Library, Duke University

Session Abstract: When the Web emerged, many librarians saw it as the perfect technology for information resources, the mythical "seamless interface" in which to search multiple databases simultaneously. Now that we have experience, what is the reality of its integration into the library world? The panel will present recent user behavior studies and their implications for adaptive interfaces, security configurations, and flexible approaches to public access computer/information resources. (From the Final Program)

Session Report:
User Behavior in the Electronic Library: The Case for Adaptive and Flexible Interfaces
by Oya Y. Rieger, Cornell University Library

Mann Library, part of the Cornell University Library, developed the Mann Library Gateway, a set of Web pages that can guide the user to many electronic resources, including both proprietary and free resources (http://www.mannlib.cornell.edu). At the time of this study, there were over 600 resources, and now there are nearly 1000. These resources could be full-text, bibliographic, numeric and spatial, or graphical. The resources are grouped by broad subject categories, and a search box is provided to search the descriptions of the databases. Oya Rieger presented the results of a user study intended to evaluate the use of the Gateway and to gather information about user requirements. The study was conducted between January and September 1996, by means of Mann Library staff focus groups involving 54 staff members and faculty and student surveys involving 26 faculty members and 16 students. The findings of the study were -- 65% of the faculty and 88% of the students were unaware of the number and variety of resources available through the Gateway when asked to estimate the number of resources. -- 77% of the faculty and 44% of the students used only one or two databases -- faculty followed their tried and tested methods for gathering information in areas they were unfamiliar with, relying on colleagues and known experts in these areas. E-mail was used frequently as a tool. -- very simple search strategies were used in searching bibliographic databases: 84% of the faculty and 100% of the students used only keyword searching. The other 16% of the faculty would also search for works by a known author.

A detailed account of the findings will be published in: Payette, Sandy; Rieger, Oya Y., Supporting Scholarly Inquiry: Incorporating Users in the Design of the Digital Library, Journal of Academic Librarianship, to be published in January, 1998.

Library Online Catalog Use in the Web Environment: The Experience of a Public Library, by Dan Iddings, Carnegie Library, Pittsburgh

The Carnegie Library of Pittsburgh (http://www.clpgh.org) is a 40 library consortium in Alleghany County, Pennsylvania. The Web is the common (only) user interface to electronic library resources, including the online catalog (DRA). The Library has 700-750 public access machines running Netscape and about 300 staff machines. These machines have unrestricted access to everything on the World-Wide Web. Dan Iddings demonstrated the creative ways to integrate the online catalog with library-produced finding aids, and other resources on the Internet, including local sites on the Three Rivers Free-Net (owned and operated by the public library).

Using a totally Web-based approach has these benefits:
-- a consistent user interface;
-- interoperability with other Web-based resources;
-- finding tools are easily created and integrated with other resources;
-- links from the online catalog to relevant Web sites are easily created (e.g., the Three Rivers Free-Net Web site pages are cataloged)

Some issues the Carnegie Library of Pittsburgh has faced: -- workstations are popular and staff have to do some policing to insure everyone gets access -- many patrons are using a mouse for the first time and need help with this.

Where our Patrons Go in Cyberspace: Click-Stream Analysis in an Academic Research Library, Laura Cousineau and John Little, Duke University (Derek Lane and Jim Coble, co-investigators)

The data for this study was collected between January and June, 1997. At the beginning of this period, the main interface to the online catalog was Telnet, switching to a Web interface in the spring. There were a limited number of workstations which had Internet access, while other workstations had access to the online catalog or to in-house databases. A proxy server was used to capture information Internet resources being accessed, the accessing workstations, time of day, etc.

This study was designed to answer these questions:
-- How are our patrons using the Web: to access library resources? to find other academic resources? for recreational use?
-- Does recreational use crowd out academic use? Scarce resources (number of computers, number of network connections, physical space, and funding) require that the existing resources be targeted appropriately. The library has no desire to censor access to the Internet.

Over a period of six months, the proxy server recorded every click at 33 selected workstations in four area: the main library (near the reference desk), the public documents department, Lilly Library and the Chemistry Library. The busy periods for each of the four locations were selected for analysis. (Graphic images were not included in counts of files accessed.)

For this study, a library resource was defined as any server whose name ended in ělib.duke.eduî or any server for subscription electronic resources. Other URLs had to be manually classified as academic or not academic. A Web site was considered to be academic if that site contained information which was similar to information contained in resources currently collected by the reference department.

130,000 site selections were made during the periods selected for analysis
7,000 distinct server addresses were accessed
The 100 most frequently selected domain names represent 30% of the selections
Library resources represented 30% of the 130,000 selections
73% of the selections were Academic
25% of the selections were not Academic
2% were Unknown

Top Eight Academic Sites:
Academic Web Sites Results
# of Hits page info
ducatalog.lib.duke.edu 25006 online catalog
www.lib.duke.edu 9835 library home page
home.netscape.com 6978
sbweb2.med.iacnet.com 5682 Infotrac on the Web
www.yahoo.com 2392
www.duke.edu 1949 University Home Page
frontier.lib.duke.edu 1925 ERL and e-reserves server
search.yahoo.com 1691

Top Eight Non-Academic Sites:
Non-Academic Web Sites Results
# of Hits page info
espnet.sportszone.com 1809
www.smokers.com 678
www.goduke.com 457
www.chelseafc.co.uk 430 Soccer Club
www.rent.net 267 Web Ticketron
countrysong.com 257
games-espnet.sportszone.com 235
members.aol.com 198

Several reference librarians noted that sites which appear to be recreational (or even pornographic) in nature could be accessed as part of legitimate academic studies.