Measuring serials usage using faculty cited journals data

Citation

Klassen, T.W. (2001), "Measuring serials usage using faculty cited journals data", The Bottom Line, Vol. 14 No. 1. https://doi.org/10.1108/bl.2001.17014aab.012

Measuring serials usage using faculty cited journals data

Measuring serials usage using faculty cited journals data

Keywords: Journal publishing, Serials, Library users, User studies

Abstract: Wesleyan University Library is refocusing the serials collection to coincide with actual resource usage, as opposed to the traditional emphasis on collecting and maintaining runs of material with the anticipation of usage. To measure resource usage, the Library has undertaken several usage studies. The primary tool is a reshelving study. In addition, to supplement this data, also used is the Institute of Scientific Information's Citation Indexes: Science Citation Index, Social Sciences Citation Index, and Arts and Humanities Citation Index, to determine which journals are cited in Wesleyan faculty research articles. The data will be used to determine which purchased serials no longer meet the needs of the Wesleyan community, as well as if there are heavily cited resources that are not currently available on campus. Detailed here are the methods used to gather the data, the techniques used to process the data, and problems with this methodology, with a discussion on the success and failures encountered measuring usage in the Sciences, the Social Sciences, and the Arts and Humanities. The results are analyzed and discussed along with a comparison with the initial data from our reshelving study.

The project and our former journal policy

Wesleyan University is a liberal arts college with 2,600 undergraduates and 200 graduate students (the Sciences and Music). The library has a collection of 2,700 current serials. Since the 1970s, the University has funded the Library at a level that has allowed it to maintain this collection without having to cancel journals due to serials inflation. The journals are organized and funded by discipline line and faculty had control over which journals were received on their line. New faculty's serial needs have been supported through canceling titles within their discipline and replacing them with newly requested titles. The faculty in each department makes cancellation decisions. The process must be dollar neutral. However, it has been our experience that faculty were reluctant to cancel any of the long runs we have, even if they are journals that are not presently used by either faculty or students. Usually, this is justified under assumptions that the Library should:

• •

  preserve long runs;

• •

  it is a "journal that we should have"; and

• •

  it is a journal possibly useful to future hires.

Of much relevance is the concern for not creating needless collection gaps. Although all of these reasons have a degree of validity, the result has been that new hires have often not received the journals they were promised when hired.

Now: our new journal purchasing and selection policy

In 1997, the Library Director decided that we needed to change the way the library allocated serials allocations based on a number of factors:

• •

  We are not meeting the needs of new faculty.

• •

  The University will add 20 (approximately a 10 per cent increase) new faculty positions over the next seven years with no increase in the libraries materials budget to support these positions.

• •

  We are expecting a significant portion of the University faculty to retire and be replaced in the upcoming ten years.

• •

  There is a perception on part of the University administration that many of the serials in the library are not being used and that serials are taking up too much of the library's budget.

• •

  The strict discipline-based division of the collection does not allow for changes in the strength and size of departments or for multidisciplinary work.

• •

  The availability of rapid document delivery through commercial vendors and increases in the speed of interlibrary loan indicate we need access to journals, not subscriptions.

• •

  The rise of e-journals, in particular JSTOR, makes the provision of paper back-files less important.

After much discussion with the faculty, a number of policies were developed for the serials collection including the following:

• •

  Retention of titles is to be based on usage of the journals rather than on maintaining runs or keeping journals solely because they are believed to be "important" journals with the possibility of future usage.

• •

  Fund lines will no longer be discipline-based. Saved revenue from all cancellations will go into a central fund for document delivery and for acquiring new titles.

• •

  Low-use journals will be provided via document delivery.

• •

  New journals will be provided in electronic form, if available, unless there is a demonstrable reason to choose print over electronic form.

• •

  New journals will be purchased from a central fund based on recommendations from faculty and a set of usage criteria. The final decision on new purchases is to be made by the Library's collection committee.

An on-going project for the library will be the need to create a residual of funds tied to this new policy, solely for new journal purchases.

Usage studies

In order to provide a case as ironclad as possible for particular journals not being used, it was decided to use more than just the traditional reshelving study to measure the usage of the library's serials. The following methods would be used to collect information about each title:

Reshelving study

This is defined as the removal of bound or current issues from the shelves. To collect this data we used a variety of methods:

• •

  Current periodicals – each title was given a barcode and every time an issue was shelved the barcode was scanned using a portable barcode scanner.

• •

  Bound non-circulating science journals – similar to current periodicals, we gave each title a barcode and placed this barcode on the shelving at the end of the run. Each time an item was shelved it was scanned. We could not use the method described below for circulating journals, as many of our science journals have not yet been bar-coded.

• •

  Circulating journals – we used the circulation function within the SIRSI system and scanned each journal that was either returned or left for reshelving.

Availability through document delivery

The Library uses Carl Uncover and Reveal as our end-user faculty current awareness service. It is also used for end-user faculty document delivery. We searched our current serial list against the Uncover database to determine which of our titles are available via fax.

Citation information

We ran searches on the three ISI citation databases to retrieve all articles published by Wesleyan faculty in the journals covered by the ISI citation databases using the address field. These records were then downloaded and the lists of cites were stripped from the records and placed into Microsoft Word. The list of cites was processed in order to determine how many times our faculty cited each journal the library owns.

Where we collected our citation data

• •

  Social Science Citation Index 1991-1998 – 470 records – 3,612 cites CD-ROM;

• •

  Science Citation Index 1994-1998 – 511 records – 13,601 cites CD-ROM;

• •

  Arts and Humanities Citation Index 1991-1998 – 569 records – 1,593 cites OCLC's Epic.

Our procedures for manipulating the citation data

1. 1.

   Databases were searched using the address field for Wesleyan University.

2. 2.

   Records were downloaded from the citation databases using the full bibliographic record, including the citations.

3. 3.

   The records were then imported into Microsoft Word 98 (Macintosh). The citations were stripped out of each record, leaving several hundred pages of cited references for each of the three Indexes.

4. 4.

   The data was then all selected using the Select All command and turned into a table using the Convert Text to Table command (see Figure 1).

5. 5.

   The first three columns were then deleted as this information (author and date of publication) was of no interest to this study. Unfortunately, due to varying lengths in journal name abbreviations, and the number of columns the names occupied, it was not possible to truncate the entries.

6. 6.

   The table was then converted back into text using the Convert Table to Text command.

7. 7.

   The data was then sorted alphabetically and the following deletions wwre made:

8. 8.
   • •

     UNPUB.

   • •

     IN-PRESS.

   Figure 1 Convert Text to Table command

   The Find and Replace functions were used to find all occurrences of J- and replaced them with Journal- which allowed for a more accurate alphabetical sort. Finally, the data was printed out in two columns and the number of occurrences of each title was counted.

9. 9.

   The final step was to enter the data into a list of all of the libraries current subscriptions. This was done by scanning the paper lists alphabetically and entering the data into a table. This was a long and tedious process with several areas where error could creep in:

• •

  Journal abbreviations versus full titles in our current title database made comparing the data difficult. Titles not recognized were looked up in the Gale Journal Title Abbreviations directory.

• •

  Our list often used the corporate author for the title rather than the strict title as used in the ISI citations.

The following tables (Tables I to VI) provide summaries of the study's results.

Findings: benefits

The benefits of using faculty citation data, from a business management perspective, are how well the procedures:

• •

  give you a long-term window of which journals faculty use without having to spend years and a great deal of money collecting shelving data;

• •

  after several years of collecting shelving data, the citation data can continue to be part of a well-balanced resource usage measurement program and can be updated with a minimum of effort;

• •

  deal with the inadequacies of a shelving study.

However:

• •

  They only cover a short period of time and are thus not an accurate representation of usage.

• •

  Many journals are only used every couple of years when a particular course is taught. Usage is then often quite heavy.

• •

  Faculty may be on sabbatical when the usage study is done.

• •

  The faculty member gets their own subscription.

• •

  Faculty say they reshelve the journals themselves and are "damned if they are going to stop"!

• •

  Problems with scanning. Analysis of our shelving data indicates that there have been many times that the students have not used the scanner when reshelving journals. Policing the scanning thoroughly would be very intensive in staff time. We have also had a number of problems with faulty equipment, lost several months worth of data and have spent many valuable staff hours trying to rectify technical issues.

Findings: problems

Our admitted problems in using faculty citation data are:

• •

  Coverage in the humanities and some sciences are not very good. For example, monograph citations are not included.

• •

  The data is weak for computer sciences, as it misses proceedings.

• •

  A lot of data processing is needed if you do it on your own rather than ordering it from ISI. However, once methods are developed, the data management is considerably streamlined. Future runs of the data can be done with greater efficiency and fewer staff resources.

• •

  The data does not catch journals that are being used by recent hires, but new journals are easily removed from lists.

• •

  The data can be expensive if you do not already own it. For instance, we only used four years of science data due to cost and availability.

Using our results

The library is still in the preliminary stage of using its serials study data. We have just been through the first round of cuts to provide an account for future new journals growth. We proposed 121 journals based on the following criteria:

• •

  Faxable through Uncover.

• •

  Neither cited nor used.

• •

  Costs more than $100.00.

The list of 121 journals was then checked against the Uncover database to determine if the coverage was adequate. Twenty four journals were determined to have inadequate Uncover coverage (20 per cent or more of the issues missing) and were removed from the cancellation list. Our faculty was given our cancellation list and each proposed to cancel the journals, which were demonstrably not being used. Given the short span of our initial usage data (about four months), there were some problems with our list and, at least, half the list has been successfully defended by the faculty. At this point, we are being very flexible on this round of cancellations. Our purpose is to demonstrate the concept.

The problem with missing issues in Uncover has caused some concern on the part of both librarians and faculty at Wesleyan University as we use Uncover Reveal as our Current Awareness Service, as well as using Uncover for document delivery. To deal with the perceived problems with Uncover Reveal, we have initiated a yearlong study to determine the actual extent of the problem. Recent advances in interlibrary loan software and procedures (Ariel, ILLiad) have alleviated our concerns about the document delivery portion of Uncover. It remains to be seen, however, if the advances have sufficiently lowered the cost of interlibrary loan to the point where it costs less than document delivery through Uncover.

Conclusions

The Wesleyan faculty has had a very positive reaction to our first round of proposed cuts. The faculty seems to have found our arguments compelling and have thus been very cooperative when we have discussed our proposals with them. They seem to have been impressed with the thoroughness of our approach. A lot of interest has been expressed in our citation study with some humanities faculty even expressing an interest in adding their monographic publications to the data. Contrasted with previous experience at other institutions where journals cancellation projects were based on hard fiscal goals with the faculty making all of the choices, this has been a far more pleasant experience. Though, of course, Wesleyan is rather unique in having the level of administrative fiscal support for our serial collection.

The citation study was responsible for identifying as used 6 per cent of the total collection or 10 per cent of the total journals that were used. The study was particularly useful in the sciences, catching 12 per cent of the science journals that were not identified by the usage study or 16 per cent of the total science journals used. The results were less significant for the other divisions with only about 4 per cent of the journals being only cited and not used. However, even with this low percentage, it did distinguish 71 journals in these divisions, that if included in our cancellation proposal might have elicited controversy.

In conclusion, the citation study added significantly to the perception of the fairness and credibility of the library's approach to canceling unused journals. It was also particularly successful in identifying journals not used during the usage study period in the sciences. However, it was not as helpful in the arts and humanities, nor, surprisingly, in the social sciences. Further analysis of the effectiveness of using citation data needs to take place when more usage data has been collected from the reshelving study.

The author would like to acknowledge the contributions of Robert Adams, University Librarian, who initiated the Library's Serials Review process, and Helen Aiello, Serials Librarian, and Rachel Cheng, Associate University Librarian, who planned and implemented the shelving usage study.

Timothy W. KlassenScience Librarian, Wesleyan University Science Library, Middletown, Connecticut, USA

Further Reading

Kreider, J. (1999), "The correlation of local citation data with citation data from journal citation reports: ranking journals at the University of British Columbia", <Library Resources & Technical Services, Vol. 43 No. 2, pp. 67-77.Hughes, J.A. (1995), "Use of faculty publication lists and ISI citation data to identify a core list of journals with local importance: 50 molecular and cellular biology publications ranked for Pennsylvania State University", Library Acquisitions, Vol. 19, pp. 403-13.Swigger, B.K. and Wilkes, A.W. (1991). "The use of citation data to evaluate serials subscriptions in an academic library at Texas Women's University", Serials Review, Vol. 7 No. 2, pp. 41-6.