Hello, I've delved into library science research concerning OPAC usage
and it seems this material could answer some of Jim's questions.
Jim wrote:
"sort of imply that people will look for groups of materials. So, either
we must assume in your case e.g. that 93% of the searches were
successful and retrieved what people wanted (i.e. 7% of heading searches
coming from a click inside the bib record would assume that people
wanted something more). Or, that the users did not click on the headings
and just did an entirely new search. In the first case (93% success
rate), this does not ring at all true to me since searching catalogs is
so difficult, and the second case (making entirely new searches and not
clicking on the headings) is extremely dismaying."
and
"The fact that the public uses browse more than staff does not
particularly surprise me."
Studies dating back from the 80s to mid 90s (e.g., Drabenstott, Larson,
Bates) showed that anywhere from 50% to 63% of OPAC searches were
"subject" searches done using almost exclusively free-text searching
(i.e., keyword in all fields). When these searches fail (zerohit
queries are common) most users assume the library has nothing on the
subject. Furthermore, when presented with a list of controlled subject
headings (e.g., LCSH) users often reported not knowing what they were
looking at or how this could be used in searching. Finally, general
browsing is a domain novice behavior and as the user expertise in the
subject increases his/her searching becomes more and more specific
(e.g., authors, titles, etc.).
This peer-reviewed data suggests that:
1) People don't click on headings because they just don't know what it
is or what it will do for them. This is intuitive when one considers
the artificial nature of controlled vocabulary and that few Web search
engines offer such controlled subject groupings (exceptions are Yahoo!
Directories, Open Directory project)
2) Browsing is used by anyone new to a subject (staff or public) BUT
browsing is a complex dynamic process which is not well supported by
current OPACS designed to support known-item searching (i.e., a specific
book, a specific controlled subject, etc.) which may explain why staff
browse more than the public: because they know how to browse in the
OPAC.
My suggestion: make the interface to "explore" controlled vocabulary
more visual and intuitive. Attempts such as Cat-a-Cone (Hearst 1997)
come to mind but this area of interface design and testing is not well
developed or funded since (I believe) most resources are channeled to
Web information retrieval (which has little or no controlled subject
headings).
I'm developing a 3D "fly-through" interface for an LCSH organized
collection but I'm having difficulty finding a library willing to "give"
me a subset of their data (i.e., subject headings (broad to narrow
terms) and the bib records to which they have been assigned). This is
pushing me towards Amazon which freely provides an API to access their
data...sad how I want to help libraries get back to the forefront of
searching technology (OPACs were THE searching tool in the 80s) but I
might just have to help Amazon stay there. C'est la vie...la politique
;-)
Charles-Antoine Julien
Ph.D Candidate
School of Information Studies
McGill University
Received on Thu Feb 07 2008 - 15:20:01 EST