by Julia Blixrud

Background

The instructions for counting serials for ARL statistics currently are:

Questions 5-7. Serials. Report the total number of subscriptions, not titles. Include duplicate subscriptions and, to the extent possible, all government document serials even if housed in a separate documents collection. Verify the inclusion or exclusion of document serials in Question 48 of the questionnaire. Exclude monographic and publisher's series. A serial is:

A publication issued in successive parts, usually at regular intervals, and as a rule, intended to be continued indefinitely.
Serials include periodicals newspapers, annuals (reports, yearbooks, etc.) memoirs, proceedings, and transaction of societies.

The selection of subscriptions as the unit of measurement was done by ARL in the early 1980s when ARL, in order to compute unit costs, deviated from HEGIS serial counts. Since serials are generally paid for as subscriptions, to compute unit costs one must use subscriptions as the variable. Similarly, the count is for current serials, not for serials that have ceased publication. (Those items are generally bound and included in the volume count). The count assumes to some extent that libraries have serial check-in systems that could easily generate a count of current subscriptions.

The acquisition of and access to electronic serials raises a number of questions regarding counting for purposes of the ARL statistics. Some of the questions raised with ARL Program Staff include: 1) how should libraries report electronic titles that accompany a print subscription, 2) can, and, if so how should, libraries report serials acquired as part of an aggregated services (e.g., J-STOR,. InfoTrac, EBSCOHost, OCLC ECO), 3) can, and, if so how should, libraries report serials that are licensed for a limited time (e.g., Academic IDEAL, Elsevier), and 4) can or should libraries report statistics for serials that are paid for by consortia?

Changes in the cataloging rules are also being considered that will affect the definition of a serial (in general to broaden the materials that are given serial cataloging treatment) and will subsequently affect determination of a serial subscription.

Suggestions have been made by member libraries to change the way ARL counts serials. This short paper will highlight some of the issues in counting electronic serials, identify the more popular suggestions for counting, and present the ramifications for both the data collection by members and the time-series information produced by ARL to track trends in serial pricing. For those interested in history, some of the same questions about counting electronic serials were raised in the context of counting microform reproduction of serials several years ago.

Some Characteristics of Electronic Journals

All electronic journals are not created equal. Some electronic journals are created and distributed electronically as print subscriptions are, but with no printed counterpart. Issues or articles may be distributed, but these journals are often managed in a similar manner to print subscriptions and can be counted as such. Other electronic journals are page images of a printed journal and thus are a substitution for the print. In this way, they are similar to microform reproductions. These journals can be counted as duplicate subscriptions are now counted.

Some electronic journals are more than a print counterpart. They include additional material and usually links to other resources. These journals can also be counted as separate, similar to duplicate, subscriptions. Other serials may be made available in a full-text format with or without graphics. These journals may contain content of print counterparts, but in some situations do not represent the complete print journal. However, if these serials are made available as separate subscriptions, they can be counted as such, regardless of the relationship of the content to the print journal.

Some Characteristics of Electronic Journal Collections

Electronic journal collections also vary. Some collections are complete runs of serial titles, while others may be only selected issues. Content may change for each new release depending on the agreements between publisher and vendor, and it is possible that not only journal issues, but complete journal titles may be added or dropped with each new release. These collections are difficult to monitor and the provider may or may not give complete title coverage information to the subscriber.

Unit Cost Comparisons

One value of the ARL statistics during the serial pricing crisis has been the ability to generate unit cost figures to document significant price increases. Expenditures for current serials divided by current serials purchased for the past 12 years have generated a vivid picture of the rate of increase of serial subscription costs to member libraries. Even though these figures do not represent the complete academic membership of ARL, the reporting base has been large enough to validate the trends. In order to generate unit cost comparisons, the reported unit must correlate with a comparable expenditure.

Counting Serial Bibliographic Records

Some individuals have suggested ARL use bibliographic record counts for serials. Most libraries have automated systems in which a count of serial records can be run. Like title counts, however, the problem with bibliographic record counts is that it would include active and inactive titles as well as title changes, and using this count implies that all serials be cataloged. Counting records is dependent on library cataloging policy regarding record creation for multiple versions of items and current library cataloging policies for electronic serials varies widely. Some libraries create separate records for each version of the serial, while others include information on all electronic versions of a title on the record for the print serial. A bibliographic record count limits libraries from including titles to which they have web access, but for which they do not create catalog records. While many catalogers support creation of bibliographic records, others have suggested that institutions with well-developed web sites can provide just as good or better access to electronic serials as those who have cataloged them.

Aggregated Services

The purchase or lease of electronic journals through an aggregator raises another set of questions. If libraries subscribe one-by-one to the electronic serial, there is no confusion regarding counting, however, the aggregation options generally mean the library is paying one fee (i.e., subscription) that results in the acquisition of a significant number of titles. These "subscriptions" may not be comparable, depending on the package provided by the vendor. Expenditures for the complete package may be completely from the library budget, or the library may contribute to a consortial purchase. If the full amount is from a library budget, the serials (individual titles counted as subscriptions) could be included in the current serials purchased). For those shared purchases, ARL members have questioned how to apportion the titles (e.g., subscriptions) to the current serials purchased and the current serials not purchased categories. Some have suggested that a formula be applied to apportion some titles as purchased subscriptions and others as not purchased serials.

Consortial Leases and Purchases

Many of the electronic journals are being acquired through purchase or lease by consortia on behalf of a group of libraries. Similar questions to counting the aggregator model are being raised regarding the leases. Libraries want to report the total titles being acquired, but their costs are only a portion of the lease or purchase price.

Counting Ceased Electronic Journals

A recent question has been how to count ceased electronic journals. For those libraries who maintain the archives and for which the serial has a defined enumeration and chronology that lends itself to constructing a volume count, the serial could then be counted as any ceased serial.

Access Inventory

One suggestion for counting serials is to separately count those items that are part of electronic access collections. Similar to the ARL access inventory, this count would be by product name and member libraries would provided an inventory of their products and services to the ARL office, which would then supply a common total resources figure for each product. The benefits from this approach are that the same product would receive the same count for every institution. However, there may be product differentiation for electronic resources based on purchase agreements or licenses. Also, it has been suggested that some products may be held by all libraries (this hypothesis has not been tested).

Separate Variable for Electronic Journals

Some members have suggested that ARL create a separate category of counting for electronic resources and include that variable in the membership index calculation. The variable could be the total expenditures libraries make for electronic information, it could be the library total plus the investment made by the institution (if data from the current supplementary survey suggests that it is possible for libraries to obtain that information), or it could be a count of electronic resources made available by the library through its catalog and web sites. It may be useful to develop some sort of variable for electronic resources, but since counting strategies are still unsettled this would be premature.

Additional Data Gathering

An ARL SPEC survey on electronic journals in academic libraries is in development. The results may be helpful in identifying how libraries are currently handling electronic serials. A short survey on counting could be prepared and distributed to survey coordinators in advance of the upcoming survey coordinator meeting in which how to count is an agenda topic.