D-Lib Magazine
October 2002

Volume 8 Number 10

ISSN 1082-9873

Comparing Library and User Related Costs of Print and Electronic Journal Collections

A First Step Towards a Comprehensive Analysis

Carol Hansen Montgomery, Ph.D.
Dean of Libraries, Drexel University
<montgoch@drexel.edu>
<http://www.library.drexel.edu/facts/staff/dean.html>

Donald W. King
Research Professor, University of Pittsburgh
<dwking@pitt.edu>

Abstract

Drexel University's W. W. Hagerty Library received funding [1] in the Fall of 2000 from the U.S. Institute for Museum and Library Services (IMLS) to study the impact of a library's shift to electronic journals on staff and costs. The goals were to perform a comparative analysis for Drexel's library (a case study) and to develop a model for use by other libraries. The results suggest that, when all costs are considered, electronic journals are more cost effective on a per use basis. Storage space for low use bound journals is a major expense. A readership survey shows that the library's electronic collection is widely accepted and extensively used. Since there are methodological difficulties with the data available to make the analyses, this study should be viewed as a single first step to address an issue of critical importance to academic libraries.

Introduction

A paper describing plans for this project, published in D-Lib Magazine in October 2000, began with the following statement:

Much has been written about the economic impact of electronic publishing on publishers. There also has been considerable discussion of the cost of subscribing to electronic publications. This paper addresses another important organizational impact triggered by the migration to electronic journals that has heretofore received little attention in the literature: the changes in the library's operational costs associated with shifts in staffing, resources, materials, space and equipment. (Montgomery, 2000)

This statement is equally appropriate as the beginning of this article, which can be considered part two of the October 2000 D-Lib article. The burgeoning literature in the library field discusses primarily issues, problems, and local practices associated with electronic journals or visions of the future, but provides little insight into how non-subscription costs are changing in libraries. One forthcoming exception will be the results of a major Mellon Foundation funded project underway at the University of California system to look at the comparative costs of maintaining print and electronic versions of journals. See Collection Management Initiative, 2002 (CMI). The impact on users will also be explored. Limited space is a key motivator. Carol Tenopir and Donald W. King have a section on library costs in their extensive review of scholarly journal publishing (Tenopir and King, 2000, p. 379) and have also analyzed differences in scientists' use of print and electronic scholarly journals. (Tenopir and King, 2002)

Drexel University's W. W. Hagerty Library received funding in the Fall of 2000 from the U.S. Institute for Museum and Library Services (IMLS) to study this issue. The goals of the grant were to (1) perform a comparative analysis of the impact of the Drexel library's shift to electronic journals on staff and costs and (2) to develop a model for use by other libraries. Drexel was well-positioned to carry out this study because it is one of the first U.S. academic libraries to migrate to an (almost) all electronic journal collection. Beginning in mid-1998, the electronic format only was preferred whenever possible [2] so that with the 2002 renewals the library subscribes to about 8,600 unique electronic and 370 print journals. See Table 1.

Table 1

Migration to Electronic Journals

Formerly Drexel Institute of Technology, Drexel is a technologically-oriented, Research Intensive university (according to the latest Carnegie classification) with approximately 500 full-time faculty and 12,000 students. It is located in an urban area of Philadelphia that borders on the central business area. Several institutional factors converged to make this rapid transition to electronic journals possible and right for Drexel. The library had:

• Strong support from the University administration
• A large majority of computer literate users
• Networking and PC infrastructure in place
• A poor current print journal collection
• A major budget increase
• The need to support growing distance education programs
• Ready access to large print journal collections in the immediate area

Another contributing factor is the belief (misguided, some would say) on the part of the library administration that, in the electronic journal world, preservation is no longer the responsibility of many individual libraries. The only model that makes sense in the new order is for a mix of organizations with national and international scope to step forward and assume the archiving role. And, in fact, this is happening. Several viable preservation models exist today: JSTOR; society archives such as those of the American Chemical Society; and publisher-library partnerships. In addition, OCLC, the Research Libraries Group, the Digital Library Fedeation and the Council on Library and Information Resources are all stepoing up to various parts of the e-journal plate. Just as this article went to press, the International Federation of Library Associations (IFLA) and the International Publisher's Association (IPA) issued a joint agreement on archiving and preserving digital information (IFLA and IPA, 2002).

Methods

The project was designed to capture ALL costs. That is, we computed all overhead and fixed costs, including allocating staff benefits and institutional overhead as is done when budgeting for grants and contracts. Costs in the broad areas of space, systems, services, supplies and staff were collected and then allocated to either (1) unbound (current) print journals, (2) bound (older) print journals or (3) electronic journals. In compiling the data, cost allocations were made according to a number of assumptions depending upon the item measured and annualized. For example, building costs were estimated at comparable rental fees; shelving was amortized over 25 years, servers over 5 years and computer workstations over 3 years. When the appropriate allocation was not obvious, costs were allocated according to proportion of use (e.g., server space to use of computer hardware and software). All of these decisions will be documented on Drexel's IMLS web site (Montgomery, 2002).

Library staff kept log sheets of their time (in minutes) spent on journal-related activities. The tasks were identified through interviews during which staff described their important print and electronic journal-related tasks. Log sheets were then designed with the specific tasks for each staff position. After a two-week trial period, staff completed the log sheets for a total of 22 weeks (eight consecutive weeks in the Spring of 2001 and then the first week of each month through June 2002).

Staff time was collected according to library department (e.g., Systems, Technical Services) and position (e.g., Reference Librarian, Shelving Clerk). This was done to make an onerous task as easy as possible, and thereby encourage compliance. But libraries are organized differently, i.e., a function located in the Systems department at Drexel might be a Technical Services function in another library. To facilitate generalization, the log sheet tasks were combined under the functions listed in Table 2, along with the allocation of other resources.

Time from the log sheets was entered into an SAS statistical program that was used to (1) compute cost based upon the average salary of the staff member in the job classification and (2) regroup totals by the functional categories. Salaries were "loaded" with benefits, institutional overhead (negotiated government rate) and an average 15 percent non-productive time.

To obtain use indicators Drexel has maintained title-by-title re-shelving counts of both current and bound journals for four years. We receive use data from publishers and vendors for about 90 percent of electronic full-text titles, compile them monthly in a single spreadsheet and compute annual totals. We define anelectronic use as an article view, either accessing an html file or a PDF download.

The library also conducts surveys to assess user reactions to the change. The Drexel library conducts a user satisfaction survey annually and in the Spring of 2002 we carried out a readership survey with the same instrument that the second author and his collaborators used to survey researchers at the University of Tennessee, Oak Ridge National Laboratory and the American Astronomical Association (King et al., 2002).

Results

Table 2 provides a summary of the operational cost data, which was computed separately for electronic, bound print and current print journals.

Table 2

Operational Costs Summary

Space Costs

The cost of housing Drexel's print bound and unbound journals annually is $245,000. This cost includes shelving costs amortized over 25 years. The space cost is based on rental of equivalent space ($20 per square foot). The space cost for computer equipment related to electronic journals is just under $5,000; this cost is based on allocating the cost of housing computer equipment based on the portion of use of the equipment and related space attributable to electronic journals. Clearly, space cost is far greater than any other single component for print journals, but is only a minor factor for making electronic journals available.

Systems Costs

Systems costs included server and workstation hardware and software purchases, and maintenance for both the hardware and software. Since the software and hardware serves many other functions, the total costs were allocated, for the most part, by the percent of use that could be attributed to e-journals. The measurements of use primarily consisted of web logs and survey data.

As shown in Table 2 systems costs, which are just over $10,000 per year for electronic journals and less than $5,000 for print journals constitute a small percentage of the total cost of all three collections. While the total library systems costs are substantial, the systems support many other services, and the portion allocated to journals is small.

Supplies and Services Costs

Supplies and services costs were either discreet, recorded costs (e.g., binding, tattle tapes) or allocated costs (e.g., printing). These costs also do not have a major impact on Drexel's journal-related costs. We make a small ($2,000) profit from printing electronic journal copies. For print titles, binding-related expenses (about $8,000 in the bound journal category) make up nearly all the costs in this category.

Local library practices could change this picture dramatically. Libraries that maintain print equivalents of electronic journals will incur much higher binding costs. Drexel University outsources all patron photocopying so the library does not incur any photocopy expense or benefit from photocopy revenue.

The choice of methods to make electronic journal holdings known and accessible will also have a major impact on service cost. "To catalog or not to catalog" is the question. To date Drexel has chosen to not catalog electronic journals but to provide access through our local WEBRAT system that utilizes an SQL database that creates web pages listing titles, holdings and links to the journal. (McLaughlin, 2002). While this saves some OCLC fees, the primary impact is in the reduction of staff costs.

Staff Costs

Total staff costs related to electronic journals are higher than those for print: $127,000 versus $88,000. Since Drexel's collections have 8,600 electronic journals and 370 print journals this is not surprising. What is surprising is that the magnitude of the difference is not greater. This is partly because during the data collection period, the print collection was reduced from about 1,000. There were more print journals at the beginning of the data collection period, and there were significant transition costs. Still, the difference is smaller than expected.

Dean/Administration

Administering the migration to electronic journals consumes a large amount of the Library Dean's time. A conservative annual total cost is $22,000. The budget, in Drexel's case, was both increased and re-allocated. Personnel have been re-organized with corresponding revision of job descriptions and position classifications. New positions were created, funded and filled, most notably an Electronic Resources Librarian position. There were regular decisions to be made regarding licenses, and some of the package prices were so large that the Dean was involved personally in negotiations. Communications with staff, users and administrators became critical to success. See Montgomery and Sparks (2000) for a discussion of issues in managing this change.

Communications [Staff]

Library staff felt that the level of communications required in managing electronic resources was sufficiently large to warrant recording this time separately. In fact the annual total came to $8,300 for all formats. The amount of time measured as associated with communications regarding electronic journals is, in fact, greater than for print. But again, the difference, about twice as much related to electronic, is not in proportion to the number of titles involved or amount of use. In retrospect this cost should probably have been allocated to other functional areas.

Collection Development and Acquisitions

Staff costs to both select and acquire electronic journals are higher for electronic journals ($18,000 and $8,000, respectively) than for print, which had annual costs of $6,000 and $4,000, respectively. Developing an electronic journal collection is much more complex than developing a print collection. Common considerations in evaluating print subscriptions are subject discipline, cost, quality, collection balance, faculty interest, availability elsewhere, and for renewals, use. For electronic journals, prices must often be negotiated, trials arranged and competitive sources and package "deals" evaluated. Other additional factors are comparability with print content and visual quality, linking capabilities, interface design, archiving policy, length of back files, availability use statistics and access restrictions such as where (off-campus use) and to whom ("walk-in" users, libraries wanting to "borrow" via interlibrary loan). Collection development staff costs were reported to be about three times greater for electronic journals than for print, again a difference less than anticipated. One factor may have been that task definitions did not make clear how to charge time spent comparing an electronic journal to a print journal.

In the former print journal world nearly all the work of acquiring journals was managed by the serial jobber. At renewal time library staff marked up a list of holdings, making additions and deletions based on program changes and the rate of inflation compared to the library's budget. The process of acquiring electronic journals is not so simple. We now find it most efficient to purchase most publisher's and aggregator's packages directly orfrom consortia. Individual journals are purchased through the serial jobbers as print only, electronic only or "bundled" (both print and electronic) with varying prices depending on the format.

While staff costs to acquire electronic journals are more than for print, given the large difference in the collections' sizes, the difference of about 2:1 is, again, not as large as expected. It is possible that some of the cost of electronic acquisition was counted by staff as print because both were involved in preparing the vendor renewal.

Physical Handling [Bindery, Labeling and Shelving]

At $22,000 the cost of physical handling of the bound print volumes is more than twice that of these costs ($10,000) of unbound current journals. Shelving costs are approximately equal, in spite of the fact that more unbound items are re-shelved. The difference is accounted for by the labor involved in the binding processes, which is not offset by the tasks of unpacking and "tattle-taping" new issues.

Record Creation and Maintenance

The costs of record creation and maintenance for both types of journals were approximately $16,000 each, another initially surprising result. However, $6,000 of the print cost was for creating holdings records which were made necessary by a change to a new library management system; and $5,000 was for updating existing print records and database cleanup, both primarily transition costs.

Record maintenance for electronic journals is the responsibility of the full-time Electronic Resources Librarian (ERL) who maintains the WEBRAT database, adding and deleting titles as required. A half-time graduate student from Drexel's library education program assists the ERL. The ERL's other areas of responsibility are coordinating, selecting, purchasing and providing access to databases, electronic books and other electronic resources.

Drexel's decision to use a web-based listing of journals rather than catalog them is based primarily on two factors: (1) the cost (mostly labor) of keeping up with a constantly-changing collection and (2) our users appear very satisfied with the web lists. The disadvantages of this approach are that users searching the catalog can be misled when they find print journal records but not records for electronic journals. They may think we do not have a title. Also, the lists do not provide the same linking capabilities from some databases as catalog records.

The ERL and her assistant do the work at a cost of $16,000. While we do too little journal cataloging to measure the cost for Drexel, $16,000 is about half the cost of a Library Clerk according to our method of computation that takes into account benefits, non-productive time and overhead. Staff needed to catalog a dynamic collection of 13,000 journal titles and to keep the records up-to-date would be far more than a half-time clerk. Reported per volume cataloging costs are estimated to be more than $20 per title (Morris, 1992).

Many libraries are both cataloging and maintaining electronic journal web pages. After an informal online survey in the Summer of 2002, Briscoe et al. (2002) concluded that there are no clearly-defined best practices for providing access to electronic journals. Writing and maintaining the database software is also time-consuming. Another viable approach is to outsource list creation and maintenance to one of the three serials management companies recently reviewed by Duranceau (2002). All the companies have attempted to estimate staff time savings from purchasing their services. For libraries that have not written systems to create web lists, these services are a very attractive option, but it is not clear how much one of these services would benefit Drexel at this time. We are investigating purchasing only the full-text database updates and minimal catalog records (also with automatic updating). Drexel's system for creating web lists (McLaughlin, 2002) has the capability of importing print journal holdings from catalog records, a very useful feature not yet available in any of the three commercial systems.

Public Relations/Communications/Publications

This cost which annually came to $11,000 for electronic, $1,000 for current and $6,000 for unbound titles represents mostly the staff time, excluding the Dean, to communicate with users about the electronic journals. We use every venue that we can identify: the Library newsletter; electronic mail messages both to the entire Drexel community and by subject specialists to their college; talks at university and college assemblies, student groups, departmental meetings, open houses, representation during student orientations; signs in the library and other public places on campus; and vendor-hosted workshops.

Reference

The cost data indicate that helping users with electronic journals was the highest staff-related cost for either format. At Drexel, reference support for the electronic journal collection came to $37,000 a year while the reference tab for answering questions related to print journals in total was $11,000. This is a noteworthy finding and indicates the importance both of making discovery and access easier for users and of promoting the electronic journals through public relations and educational efforts.

Teaching

Instruction about electronic journals is integrated into the library's growing information literacy effort along with instruction about databases and electronic books. Most teaching is done by subject specialists who tailor presentations to their audiences. The cost of time reported for covering journal topics came to $6,000 for electronic journals, $1,000 for current journals and $600 for bound journals.

Subscription Costs

It is relatively straight-forward to add the total costs of Drexel's subscriptions to print journals and to electronic journals, divide by the number of titles and compare the average. However, this simple calculation would ignore a complex set of factors that must be considered to make a comparison between electronic and print subscription costs meaningful. A "subscription" in the electronic world is not a simple payment for the annual content of a journal title. An electronic subscription often brings with it several years of back files. And the price models and electronic content vary so radically that Drexel has found it necessary to define four electronic journal subscription types:

• Individual Subscriptions which are almost always purchased from a subscription agent (e.g., Wiley titles, specialty design arts titles).

• Publishers' Packages which may or may not be part of a consortium "deal," and are acquired by purchase through a subscription agent, the consortium or from the publisher directly (e.g., ScienceDirect, Kluwer titles).

• Aggregator Journals which come from vendors that provide access to different publishers' journals. The aggregators do not drop content, only add (so far). The collections started as full-text content and added searching (e.g., JSTOR, MUSE).

• Full-Text Database Journals that provide access to electronic journals from different publishers but do not make title or issue level access available (except ProQuest). Examples are WilsonSelect and Lexis/Nexis. Titles are added or removed regularly according to the database vendor's contracts with publishers. They often have an embargo on current issues of six months or more. There is considerable overlap among the journals in these collections and between the full-text database journals and the other two types.

Furthermore, Drexel's print and electronic journal collections are now dissimilar. The print collection is now made up of popular titles such as Time, journals for which the electronic versions do not have illustrations or have illustrations of unacceptable quality, library science journals for which Drexel has assumed archival responsibility, key journals available electronically only through full-text database vendors, and journals not available electronically. Print journals are selected individually by the subject specialist staff and reviewed annually. Most of the electronic journals are selected at the broader collection level. A large number of the electronic journals in the first two categories listed above (Individual Subscriptions and Publishers' Packages) are scientific and technical titles while only a handful of the print journals are in these subject areas.

Print subscription costs are based upon the cost of the print-only title. If a title is "bundled" — that is, a library must purchase the print in order to receive the journal electronically — then all the costs are allocated to the electronic journal. In all but a very few cases (e.g., Science) the print copy is not retained.

Costs of the electronic journals are based upon the individual subscription price or the total cost of a package. The cost of journals in full-text databases is computed to be half the cost of the database since databases are used not only to find a known article or journal, but to perform general searches for information.

Table 3 shows the subscription costs for the different journal types.

Table 3

Subscription Costs

[1] If a title is bundled (that is, print must be purchased to obtain the electronic version) we allocated all the cost to the electronic journal since that is the preferred method of access, we rarely keep the print copy.
[2] Half the cost of the databases is included here because the databases serve more functions than providing full-text of journals.

The largest total expenditure is for publishers' packages, but by far the largest cost per title is for electronic journals that are subscribed to individually. These costs are sometimes the same as, and sometimes a small percentage less than, the equivalent print journal. Small discounts apply to titles that can be purchased for electronic access only which is almost always Drexel's choice. Many of the titles purchased as individual subscriptions are scientific and technical titles. These titles are far more costly than the leisure and design arts titles that dominate the print collection.

Both the publisher and aggregator packages have a much lower per title cost than electronic titles selected and purchased individually, but the dramatic difference is the very low cost of the titles in full-text databases.

Per Use Costs

A low cost per title is an advantage only if the titles are used. We consider cost per use the most meaningful measure of a journal's value. Table 4 shows these costs for the various classes of electronic journals and for the print collection.

Table 4

Cost Per Use by Journal Type

[1] Cost of only the titles for which use data is available.
[2] Use data supplied by the publisher or vendor for publisher's packages, aggregator journals, and full-text databases. Use data for individual subscriptions is extrapolated from "click" counts. Use of print only journals is from re-shelving counts. Because in some cases vendor data was unavailable for a month or two the numbers will be analyzed further.
[3] Numbers rounded.

Print and electronic use data are not directly comparable. Even when the same journal titles are compared there are some major differences:

• Span of years covered is often different.

• A current journal use is counted each time an issue is re-shelved; a bound journal use is counted each time a bound journal consisting almost always of many individual issues is re-shelved.

• If a user looks at more than one article in a print journal, the title will only be counted once.

• If a user views an electronic journal article more than once in a session, and even within a few minutes, in most cases the title will be counted twice.

Additionally, the electronic use data received from publishers is often flawed. It is not always easy to determine what is being counted. Publishers have sometimes included abstracts in full-text view counts, and sometimes there have been gaps. The publishers provide data in many different formats so that considerable effort is required to merge the data into a form that makes meaningful comparisons possible. See Luther (2001), for example, for a discussion of this issue. In order to supplement publisher/vendor use data, Drexel programmed its database system (that generates the journal web lists) to collect "outbound clicks" on journal titles. A "use/click" is recorded each time a user links to a title from the system.

No publisher- or vendor-supplied use data is available for the individual subscription of electronic journals. In order to obtain a "guesstimate" for comparison purposes, an extrapolation was made from a year's worth of "click" counts. Clicks count only accesses of titles from web pages, not accesses by other methods or multiple article readings of a title. Print journal uses also do not measure multiple uses of a single title — or cases where users re-shelve a title themselves. However, we discourage patron re-shelving and through a survey learned that most Drexel library users are only looking for a single article when they pull a bound volume from the shelf. The situation is different, of course, for the current unbound collection which is presumably often browsed.

Although it is problematic, the data described above are the only indicators of use available. We assert that, in spite of the difficulties, large differences are meaningful. Small differences are not. There are several observable large differences:

• The operational cost per use associated with print journals is much higher than that of electronic journals.

• At $30.00 per use bound journals have by far the highest per use cost. The largest cost factor is the space associated with housing bound journals since these journals have relatively low use.

• The full-text database journals are used heavily and at $1.00 per article are exceedingly cost-effective.

The cost per use figures also suggest that the publisher and vendor packages, with cost per use slightly higher than the single titles, are cost-effective. Even though the titles in these packages are not selected individually, their ready availability has led to use that is justified financially. While the use patterns within the packages have not been analyzed, a first impression is that the use is scattered; that is, there is low use of many different titles.

Readership

To broaden the framework of electronic journal metrics (see King et al., 2002), we also performed a readership survey. The survey, conducted with 496 faculty and 342 graduate students, (response rate 20 percent and 30 percent, respectively) provided estimates regarding the amount of reading from the three library collection services, as well as, amount of reading from other sources such as personal subscriptions, articles provided by colleagues, and external web sites and archives. Reading is defined as going beyond the table of contents, title and abstract to the body of the article. Estimates were made about the extent to which faculty and students read from the library collections, as well as, the total amount of reading from the three library collection services.

The survey results indicate that the Drexel faculty averages 190 readings per year from scholarly articles, about 40 percent of which are library-provided, and graduate students average 270 readings most of which are from library sources. The library electronic collection is used by 80 percent of faculty and they average 55 readings per year for this collection. An even higher proportion of graduate students use electronic journals (90 percent); doctoral students average 160 readings per year and masters students 55 readings per year. Readings from the electronic collection constitute 50 percent of total readings at Drexel, significantly higher than observed at either the University of Tennessee (23 percent) or Oak Ridge National Laboratory (32 percent) reflecting the difference in their lower access to electronic journals.

One reason that Drexel faculty read less from electronic collections than doctoral students is that they still tend to average about three personal print subscriptions that they continue to read extensively, while doctoral students have fewer personal subscriptions. About 62 percent of undergraduate students indicated they use the electronic collection, and they are estimated to average readings of about 10 articles per year from this collection. Eighty-four percent of the faculty and student respondents preferred the library's electronic journals over print.

The amount of reading provides a further "use" metric for comparison with the publisher/vendor and Drexel web page clicks to journal titles. The total readings from the library's electronic sources are 272,000 as measured by the survey data compared with 335,000 uses as measured by counts available from about 90 percent of the publishers and vendors. The two use measures appear to be reasonably consistent, considering the weakness of each measure. Reading of the current print journals is estimated to be 33,000 compared with 15,000 issues re-shelved and 11,000 readings of the bound print journals versus about 9,000 items re-shelved. Exit surveys in other libraries have shown that a single current periodical issue involves an average of about 3.2 articles read per use and a bound volume averages just over one article read by a user. Thus the amount of readings and number of re-shelved issues and bound volumes also appear to be reasonable. With readings recorded by the survey as a measure of use, the cost per reading of the electronic collection is about $2.20 per reading, the current print collection $3.80 and the bound collection $23.50. Thus, a common reading metric of cost of all three collections shows a substantial savings with the electronic collection.

Discussion

This article provides some insights concerning the results of a fundamental paradigm shift. An electronic journal collection results in increased costs of some activities, but the advantages and decreased costs in other areas outweigh the increases. The most significant benefits of an electronic journal collection are to students and faculty. Our library surveys, combined with heavy use, demonstrate broad user acceptance. And, a recent readership survey shows that by using electronic journals, faculty and graduate students substantially broaden the range of journals read; that they access articles from many locations; and that they save a substantial amount of their time identifying, locating and obtaining needed articles. The principal downside to electronic journal collections is that the library no longer maintains older materials (corresponding to those found in the stacks) and must depend upon publishers, vendors or other entities for a permanent archive. This is a major concern to the library community that cannot be ignored. But there is considerable evidence that this issue is being addressed by the most appropriate organizations: publishers, large vendors and professional organizations, foundations, and major research libraries.

This is a case study. Drexel policies and practices impact many of the specific costs reported here in significant ways. The decision to use a web-based system to provide access to electronic journals to reduce labor costs is one. Drexel has embraced the "big deal" approach to journal subscriptions, which has had the effect of reducing subscription costs per title; and possibly, reducing cost per use. The decision to forgo cataloging has also reduced staff time considerably.

When all cost factors are included, Drexel's cost per use of bound print titles is $30 as opposed to $8.50 per use for current print subscriptions, even though the current print journals are now heavily weighted towards titles that are inexpensive popular "browsing" titles. This is more even than the most expensive category of electronic journals, which is approximately $4 per use. Cost per use of electronic titles acquired as collections ranges from about $3 (publisher's packages), to about $2 (aggregator's packages), to less than $1 for full-text databases. Full-text databases, often criticized as sources of electronic journals because of the volatility of their holdings and the often missing visuals, are an important part of the overall picture. They are used heavily and are very cost effective.

On the other hand, at $30 per use the cost of providing bound journals is high. Space costs account for 80 percent of this total. The increasingly low use combined with the high cost of the bound print journals calls into question the wisdom of continuing to store most of this collection. Preservation of our cultural heritage is not the only compelling reason for creation of reliable long-term digital archives. As Bowen (Bowen, 2000) points out in a discussion of the JSTOR project, potential cost savings on both local and national level are very large.

This analysis was taken as a first step to shed light on what we believe is a crucial issue. The data presented here are for one academic library only. It is a case study and as such can only suggest answers. We rely on journal use data that is wanting, but which is the best available now. As standards develop this data will improve and lead to more precise analyses. The transition from print to electronic publications will cause fundamental changes in scholarly communications, and the impact on libraries, their funders and their users will be great. It is important that the persons guiding these changes have the best information available to inform their decisions. We present these initial findings with the hope that it will contribute knowledge that others can use, and hopefully, will build on.

Acknowledgement

Special thanks to Katie Brady, Drexel's Electronic Resources Librarian, for organizing all the cost and use data reported here.

Notes

[1] Funded in part by a grant from the IMLS, NR-00027.

[2] Exceptions were titles that are used heavily for browsing (e.g., Science and Nature), key titles where the only access was through a full-text database, and titles where visuals were either missing or did not display satisfactorily online.

References

[Bowen]Bowen, W.G. (2001). The academic library in a digitized, commercialized age: Lessons from JSTOR. Based on Romanes Lecture, delivered at Oxford University, October 17, 2000. Retrieved August 18, 2002, from the World Wide Web: <http://www.jstor.org/about/bowen.html#11>.

[Briscoe et al.] Briscoe, G.l., Selden, K. and Nyberg, C. (2002). "Best practices in connecting to online resources." American Association of Law Libraries, 95th Annual Meeting, Orlando, FL, July 23, 2002. Retrieved August 20, 2002, from the World Wide Web: <http://lib.law.washington.edu/_cheryl/cathome.htm>.

[CMI] Collection Management Initiative. (2002). Academic initiatives. Office of the President. University of California. Retrieved August 22, 2002, from the World Wide Web: <http://www.ucop.edu/cmi>.

[Duranceau] Duranceau , E. F. (2002). "E-journal package-content tracking services." Serials Review, 28:1, 49-52.

[IFLA] IFLA and IPA. (2002). Preserving the memory of the world in perpetuity: A joint statement on the archiving and preserving of digital information. Retrieved August 22, 2002, from the World Wide Web: <http://www.ifla.org/V/press/ifla-ipa02.htm>.

[King et al.] King, D. W., Boyce, P., Montgomery, C. H. and Tenopir, C. (2002). "Library economic measures: Examples of the comparison of electronic & print journal collections and collection services." Library Trends, Winter 2002. In Press.

[Luther] Luther, J. (2001). "White paper on electronic journal usage statistics." The Serials Librarian, 41:2, 119-148.

[McLaughlin] McLaughlin, T. (2002). WEBRAT. W. W. Hagerty Library, Drexel University. Retrieved August 22, 2002,from the World Wide Web: <http://www.library.drexel.edu/facts/webrat/default.html>
<http://webmail.drexel.edu/redirect?>
<http://www.library.drexel.edu/facts/webrat/default.html>.

[Montgomery 2000] Montgomery, C. H. (2000). Measuring the impact of an electronic journal collection on library costs: A framework and preliminary observations. D-Lib Magazine, 6:10. Retrieved August 22, 2002, from the World Wide Web: <http://www.dlib.org/dlib/october00/montgomery/10montgomery.html>.

[Montgomery 2002] Montgomery, C. H. (2002). Institute for Museum and Library Services (IMLS). W. W. Hagerty Library, Drexel University. Retrieved August 22, 2002, from the World Wide Web: <http://www.library.drexel.edu/facts/imls/default.html>.

[Montgomery and Sparks] Montgomery, C. H. and Sparks, J. L. (2000). "The transition to an electronic journal collection: Managing the organizational changes." Serials Review, 26.

[Morris] Morris, D. E. (1992). "Staff time and costs for cataloging." Library Resources & Technical Services, 36:1, 79-95.

[RLG] Research Libraries Group (RLG). (2001). Attributes of a trusted digital repository. Retrieved April 25, 2002, from the World Wide Web: <http://www.rlg.org/longterm/attributes01.pdf>.

[Tenopir 2002] Tenopir, C. and King, D. W. (2002). Reading behaviour and electronic journals. Learned Publishing, 15.

[Tenopir 2000] Tenopir, C. and King, D.W. (2000). Towards electronic journals: Realities for scientists, librarians and publishers. Washington, DC: Special Libraries Association.

Top | Contents
Search | Author Index | Title Index | Back Issues
Editorial | Next Article
Home | E-mail the Editor

D-Lib Magazine Access Terms and Conditions

DOI: 10.1045/october2002-montgomery