CDL-Mid-Winter discussion (#2): Use statistics for electronic resources

From: Lynn Sipe <lsipe_at_usc.edu>
Date: Thu, 16 Feb 2006 14:40:02 -0800
To: colldv-l_at_usc.edu
From: "June K. Schmidt" <JSchmidt_at_library.msstate.edu>

Collection Development Librarians of Academic Libraries Discussion Group

Topic 2.           Use Statistics for Electronic Resources

Facilitator:        Brian Quinn,  Texas Tech University

Recorder:         Leslie Rios, University of Illinois - Urbana Champaign

  A list of questions for discussion was distributed. Problems with vendor 
consistency and reporting were discussed. Defining use can be problematic. 
Does one count searches or sessions? Aggregator packages can be problematic 
because the same database may be available from several vendors who may 
each count use differently. Some vendors are Counter-compliant and some 
not, and Counter itself still has lots of weaknesses.

The issue of how much staff time should be devoted to collecting statistics 
was discussed. Some librarians try to collect stats monthly; others, twice 
a year. Some have full time staff to collect stats, others do it 
themselves, and others rely on student help. Consortial stats are often too 
general to be of use. Consortial pricing can make it difficult to determine 
cost per use, since the actual price of a title may not be available. How 
best to present statistics was also discussed. Some liaisons may not be 
interested in raw numbers, and statistics presented in the form of charts 
or tables may get a better reception. The issue of comparability of 
statistics was discussed. Use stats must always be considered in context. 
It is difficult to compare use statistics of an engineering database with 
those of an anthropology database if the university has a large engineering 
program but only a small anthropology department.

Is it possible to trust vendors? Do they inflate use data to discourage 
librarians from canceling databases? Some vendors are good about reporting 
technical breakdowns that result in statistics gaps, others are not. 
Problems of normalizing stats to correct for unusually high use by a few 
enthusiastic researchers were the subject of discussion.

Linking metasearch applications to use was discussed. Some librarians are 
using SFX stats to fill in areas in which they do not have good use 
information. The problem of metasearch applications inflating use 
statistics was also discussed. Will metasearch capability spell the end of 
tracking database use?

Should decision to retain or cancel be based on use? The general consensus 
seemed to be that we have to base decisions on some evidence, even if it is 
flawed. Librarians may want to keep their own use statistics rather than 
rely on vendors. The ERUS project was discussed as an example.



  Submitted by:

June Schmidt
Associate Dean for Technical Services
Mitchell Memorial Library
P.O. Box 5408
Mississippi State, MS 39762
662-325-7672
Received on Fri Feb 17 2006 - 02:28:08 EST