Applying the results of transaction log analysis
| DOI | https://doi.org/10.1108/eb047887 |
| Published date | 01 February 1993 |
| Pages | 87-97 |
| Date | 01 February 1993 |
| Author | Beth Sandore |
| Subject Matter | Information & knowledge management,Library & information science |
APPLYING THE RESULTS OF
TRANSACTION LOG ANALYSIS
Beth Sandore
CONTENTS
Introduction: Applying the Results of Transaction Log
Analysis
Administrative Applications
Aggregate View of System Resource Use
Terminal Locations and Traffic Patterns
Staff Training
Public Services Applications
User Searching Patterns
Problematic Searches
Null Set Searches
Unmanageable Search Results
Subject Searching
Length of Search Session, Number of Search
Terms Used
Software and Interface Needs
System Response Time and User "Think Time"
Use of Menus and Help Screens
Instructional Needs and Methods
Technical Services Applications
Authority File and Other Types of Database
Maintenance
Cataloging and Classification Research
Collection Management Applications
Systems Development and Enhancement Applications
Test the Efficacy of Changes or Enhancements to
the System
Anticipate Evolution of System Use and Demands
Determine User Preference for Experimental
Changes
Decision-Making Tool for Networks and Consortia
End-User Feedback
IR and Cognitive Behavior
Everything You Always Wanted to Know....
INTRODUCTION: APPLYING THE RESULTS OF
TRANSACTION LOG ANALYSIS
The ability to conduct unobtrusive observation of
user searching is a potential strength of the method of
information retrieval system analysis known
as
transac-
tion log analysis (TLA). Transaction logs supply
unequivocal information about what a user typed while
searching. All other methods rely on self-reporting,
which, as Nielsen points out, is not always corroborated
by the logs.1 Regardless of where in an institution
information retrieval (IR) system evaluation takes place,
TLA is a method that enables library staff at all levels
to examine a variety of system and user-related activi-
ties that are recorded on the log. Dominick suggested
that TLA can enable the examination of three broad
categories of activity: 1) system performance and
resource utilization, 2) information retrieval perfor-
mance, and 3) user interaction with the IR system.2
This article has been divided into several sections
corresponding to functional areas in
a
library to suggest
useful applications of TLA.
This article offers the reader who may not be a
systems designer a method to examine how an IR
system is being used. Where each department in the
library may be studying different aspects of the log,
the focus may vary. The overall objective is to contrib-
ute to better system design. However, the perspective
on what constitutes "better system design" may differ
depending on whether an administrator, a reference
librarian, or a cataloger interprets a
log.
This variance
of opinion is healthy, and ultimately will contribute to
better and more open discussion of enhancements that
enable the user to search an IR system effectively. This
section suggests numerous examples and potential
applications of TLA that can enhance decisions about
resources, programs, and service throughout the library
Sandore is assistant automated services librarian,
University of Illinois at Urbana-Champaign, Urbana,
Illinois.
— APPLYING THE RESULTS OF TRANSACTION LOG ANALYSIS — ISSUE 42 – 11:2 (1993) 87
To continue reading
Request your trial