assistance would continue.
More recently, Steven Bell re-
counted his well-publicized Columbia University debate with
Sarah Watstein on the future of the reference desk. Bell advo-
cates that academic reference desks should be eliminated by
2012 and offers some standard rationales for his position: mobile
technology negates the need for a physical desk; students or
paraprofessionals can staff a desk because reference questions
are largely printer- and computer-related; and a just-in-time
model of reference service (as opposed to just-in-case desk
sitting) puts professional skills to better use elsewhere, including
classrooms, residence halls, or academic departments. His
opponent argued that the reference desk remains a powerful
symbol of the culture of academic libraries; the value of personal
service should not be underestimated; reference desks are busy
despite decreased transactions; users of complex information
tools need an intermediary; a teachable moment in person is not
equal to a teachable moment online; and modern physical
reference desks are designed to complement the versatile
learning environments that are found in academic libraries.
Neal responded to Bell's call for the elimination of the
reference desk with a rather pragmatic middle-of the road stand
and calls for just the type of examination this study hopes to
Especially with relatively few staff, a library with a busy reference desk
can't just get rid of it because of some visionary proclamation. Simi-
larly, a library with a slow reference desk can't hold on to one for the sake
of tradition, especially if time and money can allow librarians to engage
in more visionary work. Ultimately, it's up to individual libraries to
decide whether traditional reference desk duties seem more useful for
their communities, or (time and money permitting) if librarians should do
more than clearing various mechanical jams and pointing patrons to the
bathroom. Some institutional honesty and a willingness to critically
examine the nature of reference desk transactions seems a good place to
Whitson advocates a differentiated service model that dis-
cards the idea of reference service as a single desk activity and
categorizes different levels of service (directions, technical as-
sistance, information look-up, research consultation, and library
instruction) which must each be structured, staffed, supported,
and evaluated on its own terms.
In a survey of literature on
perceptions of reference services from the perspectives of
managers, librarians, and users, Rieh concludes that alternative
reference service models can best be redesigned by looking more
closely at how users are dealing with their information problems
and how they get help from reference librarians in technological
This study attempts to address the latter, based
on the questions users posed at a reference desk; it looks at the
actual cost-effectiveness of staffing a reference desk based on
librarian salaries and the nature of the questions asked.
Stetson University is a private university (approximately 2500
FTE) with a main library that holds about 336,000 volumes and
bound periodicals, as well as 376,000 federal documents. The
library subscribes to more than 100 subscription databases and
holds more than 20,000 print and electronic journal titles. The
reference desk is staffed for 68 out of 92 open hours per week,
including nights and weekends, by four full-time and two part-
time librarians, all of whom hold the MLS degree. During the
study periods, the library housed between 27 and 40 public
workstations that had Internet access and the Microsoft Office
suite of software loaded; each computer also had an individual
printer attached. The library building offers both wired and
wireless Internet access for laptops as well.
For a two-month period beginning in October 2002, Stetson
University reference librarians recorded all reference question
asked at the reference desk along with the source(s) u sed to try to
answer them. All questions were recorded that came to the
reference desk in person, by phone, or by email. Although most
questions came in person, no distinction was made in the data
collection as to the manner in which the question came to the
desk. The two-month study was repeated in spring 2003, spring
2006, and fall 2006 to account for changes in class assignments
and changes over time.
The original study design focused not on the questions asked,
but on the sources used to answer the questions. Librarians
wanted to test their hypothesis that online sources were used far
more than print reference books to answer questions. The 2002
2003 data were used in a previously published study that
showed, as librarians had suspected, that online sources were
used much more than print reference sources to answer ques-
That study suggested as an area of further research an
analysis of what students want to know in other words, what
questions are students asking and can library reference services
better address the students' information needs.
The current
study, therefore, addresses that suggestion and focuses on the
questions asked over the total eight-month study period.
To collect the data, reference librarians recorded each ques-
tion along with the source or sources used to answer the ques-
tion. The first 2 mont hs, librarians manually recorded the
questions and sources used to answer them and then the data
were transcribed into an Excel spreadsheet. During the
subsequent study periods, librarians typed the questions and
sources directly into an Excel spreadsheet at the reference desk.
Questions were necessarily paraphrased, but librarians tried to
write or enter a question as soon as possible after the question
had been asked to ensure the question was accurately repre-
sented and included all the answer sources. A data monitor
reviewed the spreadsheet daily to clean up typographical errors,
to maintain data consistency, and to assign categories to the
sources used to answer the questions.
Because the original study focused on the answer sources and
not the questions, it was decided in the study design to exclude
two transaction categories as they would not yield any mean-
ingful answer source data; therefore 2528 directional and
machine transactions, though counted in the aggrega te, were
not logged as questions. Directional questions that did not refer
to the collection (the location of the restroom, a campus building,
or an office) were excluded, along with machine transactions
(paper jams, cartridge changes, and copier problems) that did not
directly relate to an information need. Directional queries that
related to the collection (the location of a call number, the
location of a dictionary, or the locat ion of a specific format or
collection), however, were included in the question spreadsheet.
Similarly, if a hardware or software question relat ed to an in-
formation need (for example, downloading a document, add-
ressing error messages, printing documents, or performing
standard functions in the Microsoft Office suite programs), it
was included in the question spreadsheet.
Data Analysis
To facilitate data analysis in the current study, the four
spreadsheets from the four study periods were combined into
September 2008 391