Using Evidence in Practice
Evidence for Development and Enhancement of a Popular Reading Collection
in an Academic Library
Timothy Hackman
Head of Resource Sharing and Access Services
University of Maryland Libraries
College Park, Maryland, United States of America
Email: thackman@umd.edu
Kelsey Corlett-Rivera
Research Commons Librarian, Liaison to the School of
Languages, Literatures and Cultures
University of Maryland Libraries
College Park, Maryland, United States of America
Email: kcr1@umd.edu
Elizabeth Larson
Information Services
University of Maryland Libraries
College Park, Maryland, United States of America
Email: elarson3@umd.edu
Received: 14 Aug. 2014 Accepted:
27 Nov. 2014
2014 Hackman,
Corlett-Rivera, and Larson. This is an Open Access article
distributed under the terms of the Creative Commons‐Attribution‐Noncommercial‐Share Alike License 4.0 International (http://creativecommons.org/licenses/by-nc-sa/4.0/),
which permits unrestricted use, distribution, and reproduction in any medium,
provided the original work is properly attributed, not used for commercial
purposes, and, if transformed, the resulting work is redistributed under the
same or similar license to this one.
Setting
The University of Maryland is a major
public research university located in College Park, Maryland, USA. It is the
flagship institution of the University System of Maryland and offers 127
undergraduate majors and 112 graduate degrees through programs in 12 colleges
and schools. The University has a total enrollment of 36,102 (26,474
undergraduate and 9,628 graduate) and a tenured/tenure-track faculty of 1,464
(4,410 total faculty). The University of Maryland Libraries includes eight
campus libraries, the largest and most central of which is McKeldin
Library, with 1.2 million volumes in the humanities, social sciences, life
sciences, business, government documents, and East Asia collection.
In 2011, the Librarian for English and
Linguistics proposed the creation of a Popular Reading Collection in McKeldin Library, in response to frequent requests from
users for non-academic reading material and audiobooks. Because of the nature
of these materials, the librarian decided to lease, rather than buy, them, and
chose Brodart’s McNaughton plan as the best way to receive
new titles that could be returned once they were no longer popular. Under the
plan, the library receives 30 books per month (up to 330 per year) and
approximately 60 audiobooks per year. The librarian set up a selection profile
for books which identified the genres that the library did and did not wish to
receive, and selected audiobooks individually from the McNaughton catalogue.
The first monthly shipment of books and audiobooks arrived in December 2011 and
were shelved in the busy Learning Commons on the library’s second floor. The
collection was promoted heavily at first via social media, the Libraries
website, posters in McKeldin Library, and a feature
in the campus newspaper; ongoing promotion has been through inclusion in the
Libraries’ printed promotional materials and occasional website news items.
Popular Reading Collection materials can be identified through the Libraries’
ALEPH catalogue, but not through WorldCat Local.
Students, faculty, and staff at the University of Maryland can borrow items for
three weeks at a time, plus one three-week renewal.
Problem
The purpose of the Popular Reading
Collection is to provide a variety of current reading and audiobook materials
that can be continually updated to reflect our users’ changing interests.
The problem we face is how to identify
those interests and predict what will be popular with our users, so that we can
assess whether we are receiving the “right” titles from the vendor. In the
interest of efficiency, selection of new titles has been ceded to Brodart staff, who, in theory, have a better understanding
of popular publishing trends. However, they serve a variety of libraries and
user communities across the country, so they cannot predict what will be
popular among a heterogeneous group of students, faculty, and staff at one
particular university. Once items are received at the library, there is also
the problem of weeding the collection appropriately to retain the items that
are still popular and to keep it to a browse-able size.
Evidence
The primary evidence used in managing the
Popular Reading Collection is circulation data. Since May 2012, we have
exported reports from our ALEPH Integrated Library System (ILS) on a quarterly
basis, which show identifying information, format (book or audio CD), when the
item was added to the collection, the number of times the item has circulated,
and the date the item was last returned. This last piece of data was not
originally included, but as one of our goals is to keep the collection fresh,
we began including it in 2014 in order to identify items that have not circulated
recently. As our McNaughton selection profile is genre-based, it was also
necessary to capture genre information for every title in our collection. This
information is not tracked by our library system, so it is added manually by
looking up titles in GoodReads, a social media
platform for sharing and receiving book recommendations that includes
crowd-sourced genre information.
Initially, this evidence was gathered to
demonstrate the collection’s popularity to administrators and funders. In
reviewing the data, it became apparent that circulation statistics would be the
best way to identify items that were not popular (had not circulated) and
therefore should be returned to allow for new items. Popular Reading Collection
items circulate at a much higher rate than our general collections, and the
statistics show that they are providing a needed service to our users. All our
data can be viewed online at http://hdl.handle.net/1903/15567.
Secondary evidence used to manage the
collection includes questions asked through our online reference system or at
the Library Services Desk, and requests for particular titles submitted via
email.
Implementation
At first, using the circulation data for
weeding decisions was rather straightforward: every quarter, we returned those
items that had not been checked out. After nearly a year, however, there were
fewer and fewer items with zero checkouts appearing on the reports. At that
point, we began to rely on how long items had been on the shelf combined with
number of circulations (e.g. items that have been on the shelf for at a least
six months and have circulated fewer than five times). Now we also consider the
last time an item has been checked out.
In the summer of 2013, we decided to
review returned books by genre to determine whether our selection profile was
meeting users’ needs, and also by format to generate a profile for audiobooks
(which had been selected individually up to that point). Genre was manually
added to the weeded book reports, and subsequent calculations based on
circulation rates showed that we should decrease the percentage of
Mysteries/Thrillers and Westerns and increase Romance and Sci-Fi/Fantasy.
The circulation data is obtained from
ALEPH reports that are provided in Microsoft Excel. We have utilized the
built-in sorting and filtering features a great deal in determining which items
to weed. When we began carrying out more in-depth calculations and assessing
the entire collection as opposed to just the weeded books, we combined the data
using Open Refine (formerly Google Refine), which is a free and open source
tool that facilitates cleaning and organizing irregular data. In this case, the
circulation reports had been generated over two years and so column headings
and cell formats varied slightly. Open Refine also allowed us to easily combine
the quarterly circulation reports into one large table with uniform data. We
could then use that clean data to create an Access database to facilitate the addition of genre information through a
user-friendly form, and the generation of complex queries such as the
percentage of items from each genre that had circulated more than five times.
Outcome
Location had the largest impact on
circulation statistics, which increased by over 10% when the collection was
moved to a prime spot by the entrance on the first floor of McKeldin
Library, even though advertising had ceased almost entirely. Not only does
every person entering and exiting the library see the collection, it is also
next to an elevator, where users often browse while they wait. Users looking
for “the fiction section” get something much closer to what they were
expecting.
Adjustments to our profile to include more
of the popular genres and fewer or none of the genres with lower circulation
mean that 75.5% of items have circulated at least once. Popular non-fiction has
proved more popular than anticipated, while Westerns have been dropped
completely.
The popularity of the collection led to
requests for a DVD lending collection and a graphic novel collection by the
library’s student advisory committee. The evidence from the Popular Reading
Collection gave weight to these proposals. The collection has also helped us
promote other, related collections. Questions about the graphic novels in the
collection are often a jumping off point to introducing users to the other
graphic novels in our regular collections.
Reflection
Adding genre information to all 928
popular reading item records was time-consuming and, unfortunately, of limited
use in the end. After we collected the genre information, we were disheartened
to learn that Brodart does not use genre designations
in the way we had thought.
Selections are made by our account
representative from a list of titles that Brodart
believes will become popular based on past sales by the author, pre-release
publicity, and other factors. That list does not include genre information; it
is up to our representative to judge whether or not a book fits into a genre we
want. (This may explain why we ended up with a number of Christian romances in
our collection, despite the fact that we had asked specifically to exclude
Christian fiction, after we increased the percentage of romances in our
profile.) In practice, this means that the data we have collected on
circulation by genre is of limited use; we can adjust the selection profile but
have little control over what titles are actually sent based on that profile.
Working with Brodart to improve the selection profile
and process will be one important outcome of this assessment.
Future topics for investigation
include: the effects of location changes
or promotional efforts on circulation statistics; comparisons of Popular
Reading and regular Stacks items with similar call numbers; identifying an
ideal size for the collection (e.g., do circulation statistics stop growing
when the collection becomes too large to browse easily?); and circulation
statistics for various users types (faculty, graduate students,
undergraduates).