evidence-based librarianship: utilizing data from all available sources to make judicious print...
TRANSCRIPT
& Technical Services 29 (2005) 169–179
Evidence-based librarianship: Utilizing data from all
available sources to make judicious print
cancellation decisions
John GallagherT, Kathleen Bauer, Daniel M. Dollar
Yale University, Cushing/Whitney Medical Library, 333 Cedar Street, P.O. Box 208014,
New Haven, CT 06520-8014, USA
Available online 13 June 2005
Abstract
As the cost of periodicals continues to rise, libraries must consider the value of titles currently
acquired or subscribed to. At Yale University’s Cushing/Whitney Medical Library (CWML), staff
employed an evidence-based librarianship (EBL) approach that combined use data from several
disparate sources to make the best decisions regarding the cancellation of specific journals’ print
format. These best-evidence sources include the following: a 3-month usage study of 1249 current
unbound print journals; statistics about 3465 MEDLINE-indexed electronic journals accessed via
ExLibris’ linking tool SFX; statistics from the Association of Research Libraries (ARL) and the
American Association of Health Sciences Libraries; and various traditional library statistics.
D 2005 Published by Elsevier Inc.
1. Introduction
The Cushing/Whitney Medical Library (CWML) serves the information needs of
members of Yale University and the Yale-New Haven Medical Center. Like other
1464-9055/$ -
doi:10.1016/j.l
T Correspond
E-mail add
Daniel.Dollar@
Library Collections, Acquisitions,
see front matter D 2005 Published by Elsevier Inc.
cats.2005.04.004
ing author.
resses: [email protected] (J. Gallagher)8 [email protected] (K. Bauer)8
yale.edu (D.M. Dollar).
J. Gallagher et al. / Libr. Coll. Acq. & Tech. Serv. 29 (2005) 169–179170
libraries in the science, technology, and medical fields, CWML is in the midst of a
paradigm shift. Because users of these libraries’ resources demand the advantages that
electronic access offers, libraries are subscribing to e-journals. Library users appreciate
the accessibility and convenience that e-journals provide them. E-journal licenses allow
libraries to provide patrons with access to journals’ current content and any available
backfiles as well.
Because the transition to the electronic environment has been so rapid, many libraries have
found themselves unprepared to manage the change effectively. The transition to electronic
access impacts many facets of the library, as different resources are required to manage the
electronic collection than are needed to manage the print collection; for instance, the costs
associated with supporting print are vastly different than those involved in managing
electronic resources. A study of Drexel University’s W.W. Hagerty Library that analyzed all
the operational costs necessary to support each format concluded that electronic journals are
significantly more cost-effective on a per use basis [1]. Libraries’ serial expenditures,
however, continue to spiral upward because of the increased cost of journals and the increased
access to content. Libraries attempting to retain control of their budgets are concerned about
the value of the services they are providing their patrons and about how efficiently they are
utilizing library resources to provide these services.
For service and budgetary reasons, the CWML decided it necessary to analyze how the
journal collection was being used. Because e-journal usage statistics provided by journal
publishers cannot be compared on a one-to-one basis with in-house print journal usage
statistics, the principles of evidence-based librarianship (EBL) were utilized. bEBL employs
the best available evidence based upon library science research to arrive at sound decisions
about solving practical problems in librarianshipQ [2]. CWML staff proposed analyzing
pertinent data from all available sources to determine the most effective and efficient use of
the library’s resources. Sources included the following:
! Unbound print journal usage statistics captured by a 3-month study of the library’s Current
Periodical Room conducted in October, November, and December 2003.
! SFX statistics of online journals accessed during the same time frame as the print study.
! Annual photocopying statistics for public copiers located within the library.
! Annual gate counts of patrons entering the physical library.
! Annual bound-journal shelving statistics.
! Annual association statistics, including the Association of Research Libraries (ARL) and
the American Association of Health Sciences Libraries.
2. Methodology
2.1. Current periodical usage study
At the time of this study, CWML’s Current Periodical Room actively received 1249 journal
titles in print. As each new issue arrives, it is processed by technical services staff and then
J. Gallagher et al. / Libr. Coll. Acq. & Tech. Serv. 29 (2005) 169–179 171
shelved by circulation staff in the Current Periodical Reading Room. For patron convenience,
new issues are shelved alphabetically by title. Issues are bound according to each journal
title’s binding schedule. Once bound, they are shelved alphabetically by title in the journal
stacks of the library.
Several methods to accurately track title use were considered. The selected method was by
far the least technical, yet was extremely efficient. An alphabetical list of the titles was
generated in Excel. Columns for each week of the study were added so that shelvers could
simply place a check mark each time they re-shelved an issue during a specific week.
Shelvers gathered all unbound periodicals in the library that were not shelved and placed
them in the unbound periodical pre-shelving area. Here they were arranged in alphabetical
order and recorded. Each issue of a title received one check mark to indicate one buseQ.Shelvers then immediately re-shelved the recorded issues in their correct place to prevent
them from being recounted.
During the study technical services staff delivered all newly received issues directly to the
circulation desk, instead of the unbound periodical pre-shelving area as is the normal
workflow. This was done to eliminate the risk that these issues would be included in the usage
study, since we were only interested in counting issues re-shelved as a direct result of patron
use. Additionally, signage was posted throughout the room, requesting patrons not to re-
shelve any issues since a study was in progress. This was done to try to limit the amount of
uncounted uses.
The usage data were collected weekly and input into the Excel file. The total use of the
unbound issues was tracked. The strength of this approach was its simplicity. The staff
involved in recording usage clearly understood the process and found it easy to track use
accurately.
2.2. SFX statistics
SFX is a linking tool from ExLibris that makes it convenient for patrons to access an
electronic resource. The Yale SFX Server maintains holdings information for all Yale
Library journal holdings. When patrons search library research databases (such as
MEDLINE, EMBASE, etc.) for specific articles, a link to SFX is provided for each
search result. If the full text of the article is available electronically, SFX will use the
OpenURL standard to construct a dynamic link to the full text. If the article is not available
electronically, SFX will prompt the user to search the Library’s Catalog ORBIS to
determine if the library system subscribes to the journal in print. Since it saves a
considerable amount of time and effort, SFX has grown popular with patrons and its use
has grown 80% each year since its inception.
SFX statistics provide request data about all the electronic journals Yale University Library
System (YULS) subscribes to that are accessed via SFX. In order for the SFX statistics
analyzed to reflect the subject areas covered by CWML’s print journal collection, only those
journals indexed in MEDLINE were extracted. At the time of this study the number of
MEDLINE-indexed titles accessed via SFX was 3465. SFX statistics were gathered for the 3-
month period that print usage statistics were gathered.
J. Gallagher et al. / Libr. Coll. Acq. & Tech. Serv. 29 (2005) 169–179172
3. Limits of the study
The shortcomings of traditional journal usage statistics have been well documented [3].
Re-shelving statistics cannot tell us how the journal was used while off the shelf. Did a patron
read an article in the issue? Did the patron read more than one article? Perhaps the patron
merely browsed the table of contents or used the journal to block an irksome air-conditioning
floor register that was blowing cold air on her. These are just some of the possibilities that we
categorize as duseT when we rely on re-shelving statistics.
Electronic journal use statistics pose similar problems. Publishers and vendors provide
reports that detail how many uses each particular title received, but again these statistics
have both technical and practical shortcomings. Other than an ethical obligation for
example, what motivation do publishers have to provide statistics that may negatively
impact libraries’ perception of the value of certain titles? In 2001 Luther wrote bPublishersare concerned that the data they share with libraries lack context. If, in the absence of such
a context, usage data seem low, the publishers fear that librarians may use such information
as a basis for canceling subscriptionsQ [4]. As for accuracy, what constitutes an electronic
buseQ for one vendor may not be the same as how another vendor defines it. Some vendors
link to the article level, while others link to the title, volume, or issue level, but whether the
user actually read or even browsed the article cannot be established. Publishers and
librarians have taken a positive step towards solving the problems currently associated with
electronic journal usage data through the creation of Project Counter, an initiative formed in
2002 to attempt to standardize the way online journal usage statistics are gathered and
shared [5].
Unfortunately, not all vendors permit nor can facilitate SFX linking within their
databases. As a result, approximately 86% of the Medical Library’s e-journals can be linked
to via SFX. Additionally, SFX is only one means of accessing e-journal articles. It
represents articles accessed as a result of an active search in a database. There are
numerous other ways a patron can arrive at the full text of an article, such as following
links through the e-journal database or accessing them using a bookmark they created to a
specific title. To better determine what portion of e-journal use SFX statistics actually
represents, a random sample of 75 titles was assessed to compare SFX and publisher
statistics. The results of this study were not entirely clear-cut. For the 10% of this sample
representing the most popular journals (as measured by vendor usage data) SFX use
comprised only 3.5% of all electronic use. For the other 90% of the electronic journals in
the sample, SFX use represented 15.0% of all electronic journal usage. In fact, for the least
popular journals SFX usage approached 60% of overall use. But overall, SFX use
represented approximately 8.5% of all electronic journal usage.
Finally, whereas the usage study of the current periodical collection only measured use
of articles contained in issues not yet bound, SFX statistics reflected use of articles from
backfiles, and therefore focused on a broader timeframe. Generally researchers studying
the subjects of science, technology, and medicine are primarily concerned with the most
recent research, but nevertheless it is difficult to quantify exactly what effect this has on
statistics.
J. Gallagher et al. / Libr. Coll. Acq. & Tech. Serv. 29 (2005) 169–179 173
4. Results
CWML has in recent years responded to patrons’ requests for convenient access to
resources by aggressively acquiring electronic resources. In FY 1999 CWML subscribed to
528 biomedical e-journals, but within 5 years this figure rose to 3391 to reflect a 642%
increase (Fig. 1). Statistics from the American Association of Health Sciences Libraries
(AAHSL) from 1998 to 2002 indicate that this trend is not unique to the CWML [6]. AAHSL
member libraries have been reducing the quantity of print titles they subscribe to while
purchasing access to an increasing number of online titles (Fig. 2).
Undoubtedly electronic access to journals and resources constitutes a major service
enhancement for patrons. However, providing remote access to journals is costly for libraries.
Data from the Association of Research Libraries (ARL) from FY 1999 to FY 2002 show that
spending on electronic serials has increased by almost 125% [7] (Fig. 3). At CWML between
FY 1999 and FY 2003 serial expenditure rose by 71% (Fig. 4).
CWML gate counts show that there has been a steady decline in the number of patrons
entering the library in recent years. Gate count figures accrue each time the exit gate of the
library is opened. Since it is a fact that oftentimes patrons will hold the gate open for each
other, the figures do not represent the actual amount of people who were in the library during
the year. However, such behavior is consistent and therefore the figures provide us a reliable
gauge to measure the percentage of decrease. Between July 1, 1998, and June 30, 2003 (a 4-
year period), a 32% decline in patrons entering the physical library was observed (Fig. 5).
As a result of the decrease in the number of patrons coming to the library, use of print
journals within the library has been decreasing. There are several methods to track this but the
most accurate and telling gauge of this decrease has been the considerable drop in the number
of photocopies made in the library’s public photocopying room. This room is located on the
same level as the bound print journal collection. It contains 14 digital photocopiers that are
Fig. 1. CWML journal subscriptions by format.
Fig. 2. AASHL average journal subscriptions by format.
J. Gallagher et al. / Libr. Coll. Acq. & Tech. Serv. 29 (2005) 169–179174
operated and maintained by an external vendor. Other than enhancements, no negative
changes that would impact patrons have been made. The cost to photocopy materials has
remained the same during the study period. Nevertheless, statistics indicate that 56% fewer
copies were made over the 5-year time period (Fig. 6). It is highly probable that use of the
print journal collection shows a similar trend.
The most notable finding of the current print journal usage study was that 53% of the
collection (657 titles) received no use during the 3 months that the study was conducted
(Table 1). While it was suspected that print use had been declining, that it was so severe was
Fig. 3. —n— Average ARL member library electronic resources expenditure.
Fig. 4. —x— Annual CWML serial expenditures.
J. Gallagher et al. / Libr. Coll. Acq. & Tech. Serv. 29 (2005) 169–179 175
surprising. Another notable statistic is that only 7.1% (89 titles) was used more than once a
month, and only 1.28% was used one or more times a week. Only a small portion of the
current print collection is used significantly. In fact 10% (125 titles) of the collection
represents 60.7% of the total use of the current print collection.
SFX statistics on the usage of 3465 MEDLINE indexed titles show that 14.8% (513 titles)
were not accessed at all during the second quarter of FY 2004. 10% of the journals
represented 56.8% of all SFX usage (Fig. 7).
5. Discussion
The one overriding observation of the study was that print journals are used only a fraction
as often as their electronic counterparts. Even though SFX represents only a small portion of a
title’s online use, electronic use via SFX far exceeds statistics for the title’s print use. For
Fig. 5. —x— CWML physical library annual gate counts.
Fig. 6. —x— CWML Copico annual photocopying statistics.
J. Gallagher et al. / Libr. Coll. Acq. & Tech. Serv. 29 (2005) 169–179176
instance, the most-used print journal during the 3 months of the study was JAMA. The print
format received 69 uses over 3 months, while it was accessed 570 times via SFX. Similarly,
the Lancet received 35 print uses versus 524 SFX hits. When compared to publisher statistics
the difference is magnified. For example, publisher statistics for the Lancet show that it was
accessed 3468 times during the 3 months of the study. It would be foolhardy to conclude that
this confirms that the Lancet’s electronic format is used 100 times more often than the print
version (i.e., 3468 hits versus 35 print uses), but undoubtedly the difference between the two
formats is incontrovertible and staggering.
The study also showed that titles that were used most in print were also the titles most used
electronically. An analysis of the top 10 titles used in each format shows that these titles
tended to be titles that included articles from a broad spectrum of disciplines as opposed to
those that are discipline specific (Table 2).
That such a large portion of the print collection was unused during the study merits
action. Maintaining the print collection is far more labor intensive and costly than
maintaining an electronic collection [8]. The resources expended by the library to support
the 53% of the current print collection that was not used seriously impacts efficiency. While
some of these resources will most likely be used at some point in time, the value of
acquiring low and no-use journals in print format is minimal. Regardless of the actual cost
of these print items, the cost of the staff time involved in acquiring, receiving, and shelving
these resources could be used more productively on other endeavors. Other costs such as
the cost of shelving space, processing space, etc., must also be considered. Most
importantly, spending money to support unused resources results in more valuable
resources not being acquired for users, and this directly impacts how they perceive the
value of the library to them.
Table 1
Use per title of current periodical (print format) for Q2 FY 2004 CWML
Number of uses 0 1–5 6–10 11–15 16+
Number of titles 657 528 38 14 12
Fig. 7. Use of e-journals accessed via SFX for Q2 FY 2003 at CWML.
J. Gallagher et al. / Libr. Coll. Acq. & Tech. Serv. 29 (2005) 169–179 177
Identifying the titles that receive little or no use is only the first step towards the efficient use
of library resources. Actually transferring existing subscriptions to online only subscriptions or
canceling titles that do not provide an online equivalent is the next step. In order to cancel print
only titles, the library will need to consider carefully with feedback from relevant departments
and discipline areas whether to drop those titles. However, most of the print titles are received
in conjunction with an e-journal package. Cancellation clauses as specified by licensing terms
with publishers, vendors, and aggregators of these titles, mandate whether the print can be
cancelled or not. Many of these licenses allow for converting subscriptions to e-only for a
cost savings usually representing between 5% and 10% of the print list price. Where such
an option is available the library should avail of the opportunity to unburden itself from
the cost of supporting the print format. To that end, CWML will be transferring 212
existing print subscriptions to online only for 2005 as allowed under the licensing terms
from seven separate e-journal packages. The final transfer decisions were based on criteria
developed in cooperation with the Yale University Libraries that covered such issues as
Table 2
Most frequently used titles by format at CWML
Top 10 current print uses Top 10 current SFX uses
JAMA 69 680 New England Journal of Medicine
New England Journal of Medicine 47 659 Nature
Science 45 639 Science
Lancet 35 570 JAMA
Nature 28 524 Lancet
Consumer Reports 27 460 Journal of Biological Chemistry
American Journal of Psychiatry 24 405 Proceedings of the National
Academy of Science
Annals of Internal Medicine 23 404 Circulation
Cell 22 386 American Journal of Psychiatry
BMJ: British Medical Journal 20 320 Child Development
J. Gallagher et al. / Libr. Coll. Acq. & Tech. Serv. 29 (2005) 169–179178
archival access, reliability of the journal interface, and allowances for scholarly sharing.
Where there is no price advantage to shedding print for e-only access, the library should
choose not to process the print even though they receive it.
Another observation about the use of print and electronic formats is that a minority of the
titles subscribed to constitutes a majority of the use. The CWML collection shows that 20%
of print titles accounted for 77.8% of print use, while 20% of e-journals accounted for 73.8%
of use. This is consistent with other studies of health science libraries’ collections. For
instance, a 1998 study of the Norris Medical Library at the University of Southern California
found that 20% of titles accounted for 60% of use, regardless of format [9].
6. Further study
While the cancellation of print is often permitted, many journal package licenses prohibit
complete journal title cancellation or only allow for modest reductions in overall spending
whether in print with online or e-only. Thus, titles are bbundledQ together as part of a bBigDealQ [10]. Therefore, further analysis of the use of titles in specific packages is warranted to
determine the value of the package. Many institutions are not renewing specific packages for
budgetary reasons without data to determine what impact this will have on the resources’
users. Analyzing e-journal statistics by vendor and package will provide libraries with useful
information to better determine the true value of each package deal.
7. Conclusion
The study provides factual information about use of the CWML journal collection. It
confirms that in general CWML patrons utilize the electronic format of articles instead of the
print version. Statistics considered from various sources determined this finding and
emphasized the need for the library to take steps towards reducing serial expenditure. To
ensure that the library remains valuable to its users in an era when the Internet is most often
the first source checked for information, libraries must continue to make authoritative
resources convenient and accessible for patrons. Undoubtedly e-journals and other electronic
resources will remain the means to accomplish this mission for the foreseeable future. It is
crucial therefore for libraries to determine how best to manage these resources, since failure to
do so will result in the continued inefficient use of resources, and an underserved user-base
whose perception of the value of the library will diminish.
References
[1] Montgomery, C. H., & King, D. W. (2002). Comparing library and user related costs of print and electronic
journal collections. Retrieved October 21, 2002, from http://www.dlib.org/dlib/october02/montgomery/
10montgomery.html
J. Gallagher et al. / Libr. Coll. Acq. & Tech. Serv. 29 (2005) 169–179 179
[2] Eldredge, J. D. (2000). Evidence-based librarianship: An overview. Bulletin of the Medical Library
Association, 88(4), 289–302.
[3] Chrzastowski, T. E. (2003). Making the transition from print to electronic serial collections: A new model
for academic chemistry libraries? Journal of the American Society for Information Science and Technology,
54(12), 1141–1148.
[4] Luther, J. (2001). White paper on electronic journal usage statistics. Serials Librarian, 41(2), 118–119.
[5] COUNTER. (2002). Project COUNTER. Retrieved October 1, 2004, from http://www.projectcounter.org/
[6] Byrd, G. D., & Shedlock, J. (2003). The association of academic health sciences libraries annual statistics:
An exploratory twenty-five-year trend analysis. Journal of the Medical Library Association, 91(2), 186–202.
[7] Association of Research Libraries (2003). ARL Statistics Interactive Edition. Retrieved October 21, 2004,
from http://fisher.lib.virginia.edu/arl/
[8] Montgomery.
[9] Morse, D. H., & Clintworth, W. A. (2000). Comparing patterns of print and electronic journal use in an
academic health science library. Issues in Science and Technology Librarianship.
[10] Frazier, K. (2001). The Librarians’ Dilemma: Contemplating the Costs of the bBig DealQ . Retrieved
November 1, 2004, from http://www.dlib.org/dlib/march01/frazier/03frazier.html