ER&L 2008: E-Resource Usage Statistics & Library Assessment

Virginia Kinman
Electronic Resources Librarian
Longwood University

"Putting the Trees back in the Forest: E-Resource Usage Statistics and Library Assessment"

Adventures in data analysis
Introduction:
Longwood's library is small, with a flat acq. budget. Just implemented new Web site, Info commons, link resolver, ILLiad, staffing changes within the past 4 years.

E-Metrics Landscape:
Global:
Measurement issues:
Data sources, standards (COUNTER, right now, for vendor data), user behaviors (HTML vs PDF loading), interfaces (auto-display of FT), open access (no good way to determine access/use)

Impact Issues: (recommended readings on conference flash drive)
Federated search, link resolver, visibility, promotion (via instruction and faculty contacts), RSS feeds and search alerts (how are counted?)

Analysis Issues:
Budget decisions (common to use these data in micro-level decisions, like "this is used and this is not", but harder to take data out to the big picture level), student learning, big picture, time

Local landscape is rest of talk
-Homegrown Access database
-some arbitrary decisions about data definition matches
-vendor data only

How can we make meaningful use/comparison of apples, oranges, kumquats, and bananas?

Constant E-Metrics Data Sets:
Identified consistently available sets of data by year (most are COUNTER data types):
DB searches by FY and semester
Article downloads by FY and semester (Virginia arbitrarily defined "semester" as specific sets of months)
Database sessions by FY and semester

Able to map these data "constants" from Access database to changes implemented at the library, such as going live with the link resolver (and the priorities assigned by the link resolver to various FT providers)

Also able to match ratios of constant data sets across types of data; mentions works from the literature that study validity of various COUNTER data pieces.

E-Metrics & Library Assessment

Need to incorporate these data with other forms of library data for whole-library assessment
Sparklines (demoed this morning in OCLC Identities thing) puts data into a simple graphical form. Virginia doesn't use Sparklines itself, but uses the concept to visually represent all kinds of data from her library. Shows increase in building and circulation usage matching to start of Information Commons; changes in makeup and usage of collections and services (after implementing virtual reference, esp); impact of link resolver (big impact on ILL - made it easier via link resolver and ILLiad to submit a request and users are following that path).

Challenges:
Validity?
Worth the time? (quotes Covey on settling for "good enough data")
Urgency (can we afford not to do this?)