OPAC queries at a medium-sized academic library: a transaction log analysis.

Patron queries at a four-year comprehensive college's online public access catalog were examined via transaction logs from March 2007. Three representative days were isolated for a more detailed examination of search characteristics. The results show that library users employed an average of one to three terms in a search, did not use Boolean operators, and made use of limits one-tenth of the time. Failed queries remained problematic, as a full one-third of searches resulted in zero hits. Implications and recommendations for improvements in the online public access catalog are discussed.

**********

Many academic libraries seek to make their online public access catalogs more user-friendly and catalog searches more successful. This paper reports the results of a study conducted in March 2007 that examined transaction logs to determine if data about searching behaviors could be used to improve the catalog interface and inform plans to update the library's Web site. The author concludes with recommendations that may be applicable to other libraries.

Background

Librarians at The College of New Jersey (TCNJ) library began considering the need for changes in the online public access catalog (OPAC) interface and library Web site in the fall of 2006. Two library groups were interested in assessing these changes. The TCNJ library cataloging department's OPAC design working group wanted to improve the OPAC interface and display, and TCNJ library's Web committee wanted to create a new Web site for the library. They wanted the new Web site to give more straightforward access to the OPAC and other library resources. One cataloging librarian involved with both groups sought to address the questions raised, while incorporating research into the mechanics of human information behavior underlying the OPAC's current usage. A study assessing the transaction logs was deemed a concrete way to begin documenting patron use of the OPAC. This study was designed to respond proactively to questions likely to be raised by both library groups. Two research hypotheses were identified as needing to be tested.

* Research Hypothesis 1: Users are more likely to employ simple queries (i.e., not use complex operators and search strategies) and will not take advantage of the value-added features available in the OPAC search environment.

* Research Hypothesis 2: When searches yield zero hits, whatever the reason, users will abandon the search.

The study reported in this paper took place at TCNJ, a highly selective public residential college focusing on the undergraduate experience. Student enrollment at TCNJ is composed of approximately six thousand undergraduate students and nine hundred graduate students. Incoming undergraduate students are in the top 7 percent of their graduating high school class, with an average SAT score of 1307. The student body is primarily comprised of students who attend college directly after graduating from high school; 95 percent of first-year students live on campus, and more than half the total student population lives on campus. TCJN has approximately 950 full-time faculty and staff, and approximately fifty of them work in the library.

Literature Review

Two types of literature were investigated in preparation for the study: research studies of patron use of OPACs, including difficulties in the online environment and with searching, and research addressing undergraduates the major patron group using the TCNJ library. Literature addressing the patron use of OPACs is vast; selected representatives are discussed here. Most undergraduates were born between 1986 and 1990 and have grown up surrounded with technology. In considering the OPAC and its ideal functionality, TCNJ wanted to take into account this important user group.

Patron Use of the OPAC

Numerous studies on the libra-/ OPAC have been published since its wide-scale adoption as a replacement for the card catalog. OPAC studies published before the mid-1990s tend to focus on OPACs with MS-DOS interfaces. Although these studies give insight into the mechanics of searching behavior as a branch of human information behavior and information seeking, they do not necessarily reflect how the current generation of library patrons is approaching the tasks of formulating queries and searching. Many OPAC studies are older than the student users in today's colleges and universities. Despite the fact that these articles cannot address the immediate question of undergraduate use of modern library systems, articles about OPAC studies from the 1980s and early 1990s remain relevant and pertinent on many levels. Peters' article analyzing the transaction logs at the University of Missouri-Kansas City in the late 1980s remains strikingly applicable. (1) Observations such as "It is amazing that some OPAC users willingly spend hours learning the intricacies of software they want to use on their personal computers, but they grow impatient spending five minutes learning the basic commands and structure of an online catalog in the library" seem as relevant today as they were almost twenty years ago. (2) Peters' overall search failure rate of 40 percent is partly attributed to system design. One might safely assume that modern users would fend keyword searching via a graphic user interface (GUI) to be more straightforward and less prone to failure than command-line searching in a MS-DOS interface such as the one Peters studied.

Borgman's paper from 1996 attempts to lay to rest the card catalog design model, noting that current OPACs had not yet moved forward. (3) "The record structure, content, and primary searchable fields are drawn from card catalog design models, with the searching functions and many of the interface design characteristics are drawn from retrieval system models." (4) Borgman synthesized research showing that "people arrive at a catalog with incomplete …