The Association of Research Libraries maintains statistics on the number of reference transactions at its member institutions. Yale developed a sampling methodology in 2000 for gathering and reporting the ARL reference data. This is a basic record of the number of questions received and answered at library reference service points, mainly reference desks, excluding only those questions that are purely directional. The data reported to ARL do not shed any light on the nature or length of the reference transactions recorded. From the beginning of the sampling strategy at Yale in 2000 to the most recent year reported, 2003, annual reference transactions at Yale have declined 28.9%. This is in line with data for reference transactions at all ARL transactions. The mean number of reference transactions for all ARL institutions reporting has declined 22.0% (Association of Research Libraries 2004).
This decline in apparent use for a major service of the Yale University Library has implications for how we position reference services. It is vitally important for us now to develop a deeper understanding of what these numbers tell us about reference service at Yale. In particular, the SQI Reference group has anecdotal information that indicates:
· reference librarians handle fewer quick questions, but the questions they handle are more in-depth, time-consuming questions. This trend has been reported by other librarians at ARL institutions (Tenopir and Ennis 2001, 40-45)
· reference service provided has changed based on the nature of the increased use of electronic resources and electronic services. With fewer patrons in the physical library, reference activities now can take place at a variety of locations, not just at traditional reference desks.
· data is gathered at reference desks, but the increasing use of electronic resources means that many of our readers no longer need to come into the physical library. They may seek assistance from individual reference librarians and selectors away from the reference desk. Data on these transactions are not being captured.
The purpose of this grant is to develop a more efficient, comprehensive library wide system for collecting data on reference transactions, including the time, complexity and location. The current reference transaction counts do not give us the information we need to validate anecdotal information. In addition to the inadequacy of the data collected, the process for gathering the data is labor intensive, involving marking sheets of paper, producing counts, reporting those counts to a central person, who then enters the data in an Excel spreadsheet. The data gathered is also not currently easily accessible to reference librarians. Making good data about questions available and combining them with other use data would help the library in developing better services. For example, information about the types of questions being asked could be combined with data about use of the library (for example, web server logs) to tell us what information the reference staff should make more easily available on the library's Web site (Mudrock 2002, 155-163)
The SQI Reference Group proposes adapting a system (originally created by Scott Matheson at the Yale Law Library) that will gather the type of data we need to effectively understand reference service and plan for its future. In particular, the proposed study will gather information on the following attributes of reference transactions:
· date and time
· broad category of the content of the question (e.g., remote access, database search help, ready reference, etc.)
· date and time of completion of the transaction
· location (reference desk, librarian office, faculty office, etc.)
· mode (in person, email, phone, fax or chat)
· patron status (e.g., faculty, student, graduate student, or staff)
The data will be gathered via a web form, which will be available from any networked computer. At selected reference desks a touch screen input device will also be tested. The web form will gather the data and automatically populate an Access database table. A simple form will be created that will let individual departments generate reports on reference transactions for the month and year. Participating departments will be Social Science Libraries and Information Services, Medical Library, and SML Research Services and Collections. Letters of support from these department heads are attached. Other selected departments may also be asked to join the pilot study.
This system will be implemented in the
Spring Semester, and will run for six months.
Budget
3 Audreys (touch pad device to pilot at selected reference desks) $360.00
Systems Programming Support (20 hours/student time) $450.00
Switch boxes for 3 reference desks (use with Audrey) $200.00
Total $1,010.00
Benefits
The Yale
Law Library has been using a system similar to the one proposed since summer of
2003. By comparing the year's worth of statistics from this system to the
statistics gathered in the existing sampling program, we can see some
difference. The sampling extrapolation predicts 5061 reference transactions
from July 03 to June 04. The full count obtained with the system similar to the
proposed system was 4168 from September 03 through August 04. This difference
could have several causes, including seasonal variation, lack of complete
reporting by staff or inaccuracies in the extrapolation assumptions (since the
data provided for the extrapolations were gathered using the same system). By
analyzing transaction reporting by individual staff member over
time, a trend toward more consistent reporting can be seen.
The most important part of
the system, however, is that it allows the law library to analyze the timing
and duration of reference work in a way that the sampling program cannot. This
data allows the library to make needed reductions in reference desk coverage in
a way that has the least impact on patrons. Over time, this data will allow the
library to substantiate the theory that while reference transaction numbers are
declining, duration and complexity are increasing. The data gathered also
allowed minor changes in staffing schedules that distributed reference
questions more evenly among staff.
This study seeks to answer the
following: Are the data we collect
helpful to understanding reference activity in the library, and can it aid
departments in planning service? Sharing
the results with the participating department heads and seeking their feedback
on the usefulness of the data will measure the outcome. The departments heads will be asked
specifically what activities might be planned based on the results of the
study, and whether the results might have any use in marketing reference
services. Secondly, will reference
librarians view the collection form useful and not overly burdensome in their
workday? This will be measured by the administration of a very brief survey at
the midway point and end of the study.
Based on these results, i.e., weighing the potential benefits to the
library versus the demands on staff for gathering the data, the SQI Reference
Group will make recommendations to the SQI Council on possible changes to the
Library's data gathering practices for reference transactions.
Association of Research Libraries. ARL Statistics Interactive Edition). April 21, 2004. Internet on-line. Available from <http://fisher.lib.virginia.edu/arl/>. [October 22, 2004].
Mudrock, Theresa. 2002. Revising Ready Reference Sites: Listening to Users through Server Statistics and Query Logs. Reference & User Services Quarterly 42, no. 2: 155-163.
Tenopir, Carol, Lisa A. Ennis. 2001. Reference services in the new millennium. Online 25, no. 4: 40-45.