Report of the CDC Working Group on Serials Duplication in an Electronic Environment
July 19, 2001
The
CDC Working Group on Serials Duplication in an Electronic Environment was the
brainchild of CDC chair Will Wheeler, who suggested its formation and wrote the
context and the charge.
Context
Many
have noted the increasing duplication of serials in the electronic environment,
whether it be through buying both print and electronic versions of a title or
multiple electronic versions through more than one vendor (or both). Four easy
local examples are the multiple copies we buy of Science (12+ print copies and two online versions from Ovid and the
publisher), of the Economist (3+
print and 3 online versions via EBSCO Academic Search, Lexis-Nexis Academic
Universe, and Dow Jones Interactive), of the New York Times (8+ in print, at least 2 online via Dow Jones
Interactive and Academic Universe) and Books
in Print (12+ print copies, 1 online version via OCLC).
While
the reasons for duplication of print versions in the dispersed Yale environment
may be easy to identify, they become less certain in an electronic world. In
light of the Medical Library's decision to make changes and cancellations that
generated $43,000 in savings, the Collection Development Council believes it is
time to consider the issue more widely, even though issues of duplication have
been considered before.
The
work of the CDC News Resources Working Group taught us that we are not likely
to get rid of all duplication, but that discussion of examples can reveal
patterns and suggest collection strategies.
Charge
The
CDC Working Group on Serials Duplication in an Electronic Environment will
consider duplication among electronic resources and between electronic and
print versions of the same title, with an eye toward greater efficiency and
cost savings. While the main focus of its work will be periodical
subscriptions, it is encouraged to consider some annual and reference works
(“serials”) as well. While the group is not charged to compile a complete list
of duplicates, it may be able to estimate the number, describe patterns of such
duplication, and propose ways to achieve cost savings. Another useful task the
group will perform is to identify the reasons we buy duplicates. Enumerating
those reasons, particularly in support of services, will be useful to the
larger selector community. Finally, the group will develop some guidelines to
aid librarians considering whether to retain or cancel duplicates; that is,
some principles that will identify circumstances that argue for duplication and
circumstances that argue for cancellation when there is more than one copy and
an electronic version.
In
the Yale Library System multiple print subscriptions were purchased for the
convenience of our patrons in a decentralized library system. Librarians in
libraries collecting similar subjects, but physically distant from each other,
found it made sense to purchase a title for their location even if it was held
elsewhere on campus. Examples of such subjects include life sciences titles at
Kline and Medical and journals on environmental toxicology at EPH and Forestry.
General reference works and high-use titles of general interest were also
purchased for several locations (e.g. Science,
Wall Street Journal, etc.). With the increase in interdisciplinary
studies, many journals logically belonged in more than one library
collection. Also, since only one person
can use a print journal at a time, multiple copies of high-demand titles were
purchased within single libraries and across multiple locations.
Now
we are in a new electronic universe. Yale patrons currently have electronic
access to over 10,000 unique journal titles. Of these, at least 1,351 are
available in two or more different electronic versions. Now that we have
assembled a critical mass of electronic journals that are available 24 hours a
day, 7 days a week and can be used by multiple persons from their offices and
homes, the need for some of the duplicate print subscriptions is obviated. In
addition, libraries are facing severe space shortages in their stacks and
sending volumes off-site to the Library Shelving Facility. It is time to
revisit the issue of duplicate print subscriptions in the context of our
current electronic environment.
We
are indebted to Suzanna Lengyel, who ran reports that identified current
duplicate print subscriptions in Orbis, and to Kim Parker, who ran Suzanna’s
reports against yelmo, the Yale database of electronic journals. Suzanna’s
report looked for duplicate subscriptions with open order/pay/receipt records,
which would indicate they were current subscriptions. Although electronic
journals are listed in Orbis, it was difficult to match them with the print
duplicates. At Yale separate records are created for online journals, so they
do not appear on the same record with the print versions; in addition, the
electronic version has a different ISSN from the print journal, so titles
cannot be matched using that field.
Fortunately,
Kim was able to run Suzanna’s file against yelmo to give us an Excel file of
duplicate print subscriptions at Yale that have electronic equivalents. She
used a program designed by Wesley Most to search ISSN’s and titles in jake and
pull out the jake id’s. The jake id’s were added to Suzanna’s spreadsheet. Then
Suzanna’s file was run against yelmo and matched on jake id’s to identify print
titles with electronic equivalents. Lastly, Kim did another comparison of
Suzanna’s file and yelmo, matching on exact title (to attempt to capture some
of the titles that didn’t have jake id’s). This search only worked if the
titles were identical, however (e.g. “and” would not match with “&,” etc.).
These last two matches in yelmo extracted further information on volume extent/electronic
holdings and the source(s)/interface(s) of the electronic version(s) to include
in Suzanna’s file.
The
resulting data is only a snapshot, as new electronic journals are added daily,
but it is the best description we have of the extent of serials duplication at
Yale at this time. Kim ran Suzanna’s first report against yelmo on April 19th
and the second report on May 15th (so
that the Kluwer journals, for example, which were added in July, are not
included). These files were combined and cleaned up and given to Julie Linden
who loaded the data into an Access file.
We
created a second file of just the duplicate print subscriptions (i.e. titles
that do not have electronic equivalents), because we thought that information
might be of interest to selectors. Some subscriptions come in other formats
such as microfiche or microfilm, particularly at SML & SSL. If the print is
discarded when the microformat arrives, then that title was not treated as a
duplicate subscription. If both formats are retained in the collection, then
that title was considered to be a duplicate subscription. For the convenience
of selectors, there is a separate spreadsheet of all the titles in a
microformat which appeared in Suzanna’s report and are available electronically
(at least, all the titles which had “mic” in the location code. There are
titles at Divinity which do not have “mic” in the location code and therefore
couldn’t be sorted this way). In some cases, subscriptions which are housed in
technical services or in various librarians’ offices have also been counted. If
anyone wants to receive a copy of the large spreadsheet of print duplicates or
the spreadsheet of duplicate titles which we get in microform/fiche, please
contact Cynthia Crooker (cynthia.crooker@yale.edu).
Julie Linden then took the data from Excel and put it into an Access database, which is being made available on the CDC website. The database includes some pre-designed queries and basic instructions to selectors on how to design their own queries. Julie is willing to answer selectors' questions about the database. Selectors are reminded that the database is only a prototype; we are not planning to redesign it or update it.
Since
the database was not yet ready when most of this report was written, the
following figures were taken from the Excel file. The file of duplicates with
electronic equivalents contains 318 unique titles. Of these, SML has the
largest number of duplicates at 166 titles, followed by Medical with 121 and
SSL with 92. Kline has 85 duplicate titles (99, including ksl,ant) and Divinity
has 62. Because Orbis is inconsistent in its use of locations, when one sorts
by location, one does not get one alphabetical list, for example, of titles for
SML. Instead one gets titles for the locations sml; sml,mic; sml,per; sml,phi;
sml,refi; and sml,y.
KSL
and Medical currently have the most overlap, with 63 titles in common, while 61
titles are held at both SML and SSL.
The
estimated cost of this duplication is $153,077 (the cost of one print
subscription for each duplicated title). In some cases, there is more than one
duplicate subscription, but the cost of each title was only counted once. There
are slight variations in price among subscriptions at different libraries,
sometimes even when they use the same serials vendor, so one price has
been arbitrarily chosen. Where the
price information was not readily available from Orbis, the price in Ulrich’s International Periodicals Directory
(ulrichsweb.com) was used.
Aggregators
such as EbscoHost Academic and InfoTrac OneFile provide electronic access to
the largest numbers of duplicate titles. EbscoHost Academic is the source of
125 titles, plus an additional 16 from Ebsco Online and another 11 from EbscoHost
Business. InfoTrac OneFile is currently the second largest provider, with 94
titles. WilsonWeb contributes 41 titles, while Elsevier’s ScienceDirect is the
source of 24 duplicate titles.
On
the medical side of campus, there is a strong preference for electronic
resources. Medical Library patrons want the most current information and they
prefer to access it from their homes, offices, or the hospital, rather than
having to come to the library. Electronic journals have an inherent advantage
in their availability 24 hours a day, 7 days a week, and in their search
capabilities. One of the most difficult decisions in cancelling print is
whether the electronic version is an adequate substitute for the print version.
This is a very complicated issue which was addressed by Kimberly Parker and
CODGER in 1998. (See the Appendix for a list of the many points to be
considered in making this decision.) Another difficulty is the reaction of
faculty. It is always a good idea to solicit faculty input and/or check the
impact factors of journals before cancellation to avoid upsetting faculty.
Not
all electronic products are created equal. A distinction can be made between
collections of electronic journals, which tend to mirror their print
equivalents, and aggregator databases, which are primarily collections of
articles, may not be browsable, and may not include the full content of the
print originals. Aggregator databases may also present problems when one tries
to print graphs, tables, and illustrations (if they are included at all). The
Periodical Price Survey article from Library
Journal quoted in the next paragraph contained tables that were unusable
when printed from EbscoHost’s Academic Search Elite. (The columns didn’t line
up right, so the data had to be photocopied from the print copy of the
journal.)
Sam
Brooks, Senior Vice President of Sales & Marketing at EBSCO, sent an email
message to ebsco_web detailing the differences between full-text databases and electronic
journals. He warned that many journals in aggregator databases have embargo
periods and, furthermore, since the publishers own and control the content,
ejournals can disappear from these databases at any time. “For these reasons,” he advises, “full text
databases are not a practical, long-term substitute for print or e-journal
subscriptions purchased or licensed directly from the publisher.” According to
him, “databases should be viewed as a complement (not a replacement) to the
core print and electronic journal collections.” 1
If
an ejournal is determined to be an adequate (or superior) substitute for the
print and is affordable, some selectors may want to replace one print copy with
the electronic version. (At this point, Yale selectors are not even talking
about replacing unique print copies with electronic only access.)
Faculty
in certain disciplines are reluctant to use the electronic versions of their
favorite journals. If they are accustomed to sitting in the periodical reading
room reading journals, then they will not find the electronic version an
acceptable substitute. Almost no one reads articles online; most people print
them out. The cost of printing and the cost and inconvenience of photocopying
could also be arguments against cancelling print for some people.
Selectors
may also want to retain in print those journals with high impact factors whose
importance in their fields is undisputed, even though a good electronic
substitute exists. A professor who had hitherto been a strong proponent of
adding electronic journals at the Medical Library was dismayed when a key
journal in his field was cancelled in print. He made a case for having both
print and online access to the top journals in his discipline.
The
cost of the journal (print or online) is another important factor. According to
the Periodical Price Survey in the April 15th issue of Library Journal, in 2001 the average
cost of a humanities journal published in the United States was $146.28, the
average cost of a social sciences journal was $357.44, and the average cost of
a science journal was $799.76.2 For
journals published outside of the United States, the average cost for journals
in the humanities in 2001 was $253.68; for social sciences, $735.95, and for
science, $1211.06. Clearly there are greater budgetary constraints and greater
cost savings from journal cancellations in the sciences. It might not be
considered worthwhile to risk offending faculty by cancelling an inexpensive
humanities journal.
Finally,
there are occasions when the electronic version is unavailable, due to server
problems at the originating site or at Yale or to heavy Internet traffic. For
key journals, having more than one subscription may be advisable.
Once
selectors have determined the preferences of their constituents and the
relative importance of the various journals in their disciplines, they are
ready to evaluate the electronic equivalents available. This is perhaps the
most daunting aspect of the task of de-duping, but the Appendix below (based on
a CODGER document) provides guidelines
for evaluating electronic resources. It may be used to compare print and online
versions or multiple electronic versions of a product.
While
many libraries are cancelling unique print subscriptions in order to afford
electronic access, the Yale University Library System is in the enviable
position of considering whether or not to cancel duplicate print subscriptions.
The Working Group’s report should be seen as a snapshot of our collections in
the spring of 2001 and a prototype for future projects. It is meant to start
selectors thinking about their journal collections in an electronic environment
and to suggest what kind of data and reports to generate from the new library
management system in order to facilitate cancellation/retention decisions.
Orbis2 should alleviate some of the difficulties of gathering and combining
data about print duplicates and electronic subscriptions. In the meantime, we
offer a rudimentary Access database with which to begin exploring these issues.
The
CDC Working Group on Serials Duplication in an Electronic Environment
Cynthia
Crooker, chair
Rolfe
Gjellstad
Emily
Horning
Julie
Linden
Rochelle
Smith
A Codger document revised by
Kim Parker 8/6/98
Additions by the CDC WG are
in italics
The
issues scrutinized when deciding whether or not to duplicate are for the most
part the same issues considered when acquiring a unique resource. This document
was created because the cost-benefit analysis process takes on a critical role
in duplication, and it becomes more important not to overlook factors as a
decision is made. Please note that each resource/content may need different
weights applied to each factor.
The
question of duplication of materials is relevant in many settings of collection
development. The following list of factors is targeted to situations involving
electronic resources.
1.
Duplication
between electronic version and another format
2.
Duplication
between two different electronic versions
Consider cost. Are duplicates packaged to lower total cost? Are discounts
given for later duplication? What are the total local support costs for each
item? Do the benefits outweigh any additional costs? (Keep in mind hidden costs
such as binding, shelving, processing.)
Is there a discount for the
online if you own the print? Is the online free with print? Does the online
cost as much as the print? Cost has dictated that the Medical Library cancel a
print subscription held elsewhere on campus when the online price is the same
or greater than the print price.
Consider search/use
capabilities. Does
either item provide better capabilities for searching or using the content
(e.g. browsability)? Do the benefits outweigh any additional costs?
Consider archiving needs for
the resource. Does
ownership exist over different versions/formats? How hard (costly?) will it be
to maintain long-term access to different versions/formats?
For example,
Yale may lease one version but have purchased permanent access to (or ownership
of) another version. Maybe we've gotten
permanent access rights to more than one version, simply by virtue of generic
licenses that included those rights. The library needs to consider such
questions as whether it's cheaper to archive a print copy or a pdf or an html document and whether it’s
cheaper to do it ourselves or ask the vendor to do it, etc. It is probably
cheaper to archive print than electronic, but is it still cheaper when you
factor in the subscription costs? How
do you weigh the fact that minor portions of the content (letters to the
editor, news items) may be unavailable in one format? These are all pretty much
unanswerable questions at this point, but they need to raised. Since we will
have at least one print copy on campus, this is not as difficult an issue as it
would be if we were considering cancelling unique print subscriptions at Yale.
Consider content. Is it the same? What differs in each version?
This
is often far from easy to determine.
Consider access. How important is a particular access method for the
audience of this resource? (e.g. telnet vs. web, SGML vs. HTML, specific client
software vs. generic client software)
What format(s) do your
constituents prefer? Will there be resistance to using electronic versions?
Does some reeducation have to occur before your users will accept electronic
journals? Or, conversely, is there a strong preference for the convenience of
online access?
Consider reproduction
capabilities. How important is
printing/copying/downloading of a resource? (e.g. is the printing of an
e-resource as easy and legible as photocopying of a print resource?)
At the Medical Library,
there is a marked preference for pdf files because of the way they print
graphics.
Consider integration in a
work environment. Do different versions/formats incorporate content into a larger whole?
Is it important to provide this environment as well as distinct access to the
resource?
Consider quality of service.
Does having
either item under local control provide better quality of service? (e.g. faster
response time)
Consider user restrictions. Do different
versions/formats restrict use to a subset of our user population? Does either
item allow walk-in use? Does either item allow us to provide interlibrary loan
to other libraries?
Can you use articles in
coursepacks or electronic reserves?