STANDARDS AND METHODS FOR EVALUATING VIRTUAL REFERENCE
Looking at the Bigger Picture: An Integrated Approach to Evaluation of Chat Reference Services
M. Kathleen Kern
SUMMARY. Virtual reference offers some unique new opportunities for evaluation due to the richness of the transcripts and other automatically collected data. To put evaluation of virtual reference into context, however, libraries should view and evaluate virtual reference as part of the whole of a library's reference service. Holistic evaluation pursues an integrated approach to evaluating the total of a library's reference service. doi:10.1300/J120v46n95_07
[Article copies available for a fee from The Haworth Document Delivery Service: 1ā800-HAWORTH. E-mail address: <[email protected]> Website: <http://www.HaworthPress.com> Ā© 2006 by The Haworth Press, Inc. All rights reserved.] KEYWORDS. Evaluation, reference services, virtual reference, e-mail reference, chat reference, integration
INTRODUCTION
Virtual reference services provide the library community with a new opportunity to evaluate our services. In fact, chat reference has sparked something of a renaissance in evaluation of reference services. There are many aspects of the virtual reference transaction that can be examined; some of these types of evaluation are unique to the virtual reference environment.
Most of the evaluations of virtual reference have focused exclusively on chat and e-mail transactions, examining only those virtual transactions. As virtual reference becomes less of a novelty and more of a mainstream service, it is important that libraries start to evaluate their virtual services in the context of their reference services as a whole. Holistic evaluation will give us a better picture of our reference services.
TIMELESSNESS OF EVALUATION
The basics of reference evaluation are timeless and classic. We seek to answer the questions: who are our users and how do they use our service; when, how, and with whom should we staff our service; are our answers accurate; and how satisfied are our patrons with the assistance we provide? The fundamental questions, the reasons for evaluation, remain the same. Approach to these questions, however, has varied. Librarians have used different research methodologies and have examined different aspects of the questions.
Innovations and new technologies emerge and change the patterns of what and how we evaluate reference service. The Brandeis model of tiered staffing led to evaluations of where and with whom we staff at tiered service points, as well as the efficacy of this model. Telephone reference led to evaluation of the telephone as a communication tool for reference and a separate reference service point. Seminal papers such as Dorothy Cole's 1946 study examining the types of questions asked in public, academic, and special libraries and Joan Durrance's 1989 study of the importance of user satisfaction as a measure of reference service success led to further studies that attempt to look at the same question from a different angle.i The results of the ā55% studyā1 led to more studies of accuracy and also prompted the more qualitative evaluations of patron satisfaction and willingness to return.
The emergence of virtual reference as a popular service has had a noticeable impact on evaluation of reference. (It is not really a new technology, having been around in some form for over 20 years, but it has only recently gained widespread implementation as a library service.) A search of Library Literatureii finds six articles on reference service evaluation for the years 2002ā2003. Five of these articles were published in 2003, and only one of these is about in-person reference. None of the articles include evaluations across ātraditionalā and virtual reference services. Looking at Library and Information Science Abstracts, there were seventeen research articles indexed on reference service evaluation for 2002ā2003.iii Eleven of these were about virtual reference services. The other six only examined in-person services, with three of these articles being from Norway (there were two articles from the UK about virtual reference). Only one article2 contained evaluation of both virtual reference and in-person reference services.
Since virtual reference is new, it has turned our heads and we have focused our evaluation efforts in this direction. What is it about evaluation of virtual reference that is distinctive and what is unchanged from the evaluation of other reference services? I will explore the answers to both of these questions, as well as the importance of a holistic approach to reference evaluation that integrates evaluation of virtual reference with other reference services.
DEFINITIONS
For an article on virtual reference services, it is necessary to define a few terms since terminology may be unfamiliar. Terminology in this area is not standardized, so you may see the same words used elsewhere with variation in meaning.
Chat ReferenceāReal-time communication between two users via computer. Chat reference allows users to communicate instantaneously with librarians, or as it is commonly described, the communication is synchronous.3
E-mail ReferenceāCommunication via electronic mail. Patrons can send messages at any time to be answered by operators at another time. Since patrons are not interacting with the librarian in real-time, this mode of communication is asynchronous.
Virtual ReferenceāAn umbrella term that encompasses chat and e-mail reference as well as emerging reference communication technologies such as voice-over-IP and online videoconferencing. Virtual Reference focuses on the interaction between patron and librarian (operator) whereas Digital Reference is a broader term that includes online resources as well as virtual communications.
TranscriptāThe text of a virtual reference interaction. This may take the form of a chat transcript stored in a database of chat interactions, or the e-mail correspondence between patron and librarian. Most commonly in the virtual reference environment, the supporting software automatically collects the transcripts.
OperatorāA generic term for the person on the answering end of a virtual reference interaction. The operator may be a librarian, a paraprofessional, a graduate school student, a contracted employee from a virtual reference software company or some other person designated to answer the virtual reference questions at a library or Ask-A service.
UNIQUE ASPECTS OF EVALUATING VIRTUAL REFERENCE
Aside from the newness of these services, there are some distinctive characteristics of the virtual reference environment that make evaluation of virtual reference different from evaluation of other reference services. The most unique, and tantalizing, aspect of evaluation of virtual reference services is the availability of a transcript of the entire reference interaction. It is perhaps this aspect, as much as the newness of virtual reference, which has engendered interest in the evaluation of virtual reference. Most commercial chat software collects and archives the transcripts for either all transactions or selected chat transactions determined by the operator. These transcripts are rich with data and opportunity. The reference interview can be examined in detail as well as the accuracy and appropriateness of the librariansā answers. Tone, typing skill, and jargon are right there in print. There is a ridiculous amount of information that could be mined. It is important to look at the transcripts as the last step of the research design rather than the first; you should know what questions you want to answer before you jump into the transactions as a data source. Starting from the transactions could lead to specious evaluation. For instance, examining typing errors in chat transcripts is possible, but it has questionable value.
The consortial nature of many chat reference services can add a layer of complexity to evaluation of chat reference which is not present in the evaluation of other reference services. If you are a member of a chat consortium, you need to consider if you want to evaluate only your operators answering questions for your institution's patrons, or your operators answering the questions of other institution's patrons (or both). Or maybe you want to evaluate operators at other libraries answering questions from your patrons. There are issues of access to the consortial data, but also issues of operator privacy and inter-institutional collegiality. The drive to evaluate as a way to measure and maintain service quality can create a tension with the desire for harmony within the consortium.
WHAT IS EASIER TO MEASURE?
There is also much data that is collected automatically by the virtual reference software. The exact data will differ by vendor and the preferences of the library. Some common things that are collected: time and date of transaction, operator name, length of time spent in a chat session, user information such as affiliation or status, IP range, and Web browser. From this data a variety of reports can be run. It takes next to no effort to collect and can quickly yield much information. Some of it may be worth the time it takes to evaluate and some of it may be superfluous. Detailed reports of traffics by time of day, day of week, or week of the year can be generated to help with staffing patterns. If your patrons are asked for status or affiliation (undergraduate, public, faculty), a report can be produced to answer the timeless question of āwho are our patrons.ā If IP range is collected, you can determine where your patrons are when they ask questions (inside the library, elsewhere on campus, off-campus). These can be useful facts for training and staffing. Again, specifics vary by vendor, but this kind of patron information is often stored separately from the transcript of the reference transaction to further ensure patron privacy.
One of the ways that libraries and software vendors have made collection of patron data easy is to use a form to collect data up-front (Diagram 1).iv How difficult it is to run reports from the data in your system depends on the software, where the data is stored, and what pre-scripted reports are available or the extent to which you can write your own reports to query the database of transcripts and transaction data. Some data may be collected (such as IP), but not pre-scripted into a report or stored in such a way as to make queries of this data straightforward. Extraction of what you want t...