ARL Statistics Serials Count

ARL is changing the way it counts serials in the ARL Statistics Survey for the 2006-07 ARL Statistics survey. After many years of stability in the ARL Statistics definitions where serials were counted as subscriptions we are moving into a count of serials based on titles. In particular titles will be reported in a deduped count and with preference for the electronic format.

ARL has instructions with tracked changes available on the ARL website.

The revised instructions are placing emphasis on the content that we need to describe by focusing on titles rather than subscriptions. They are also emphasizing the electronic aspect as we move into a future where the availability of content electronically is becoming the dominant desire of our library users.

Please use this thread to ask questions, clarifications, and further the discussion on this topic.

Best regards,

Martha Kyrillidou

Director of ARL Statistics and Service Quality Programs

Association of Research Libraries

8 Responses to ARL Statistics Serials Count

  1. martha kyrillidou December 6, 2007 at 7:18 am #

    Re: question about branches and new serial counts

    The following questions was posed by a number of ARL libraries through email and I am providing a public version with the answer here:

    Q: “I attended the webcast today and had a question about the new serial
    counting procedure and branches. Are we to de-duplicate serial titles
    amongst ALL branches? In particular, I’m trying to determine what this
    will mean for Law and Health Sciences. “


    It may not be possible in some situation but the intent is to try to duplication with law and health sciences whenever possible. Some libraries are expressing a similar concern in terms of how one decides to allocate what portion of the titles to different branches but one may determine that based on local practices and procedures (i.e. are expenditures easily allocated across branches — this may make allocating titles across branches a bit easier. If not, some other shared understanding needs to be developed. Please document local practices in the footnotes.) — Best, Martha

  2. martha kyrillidou December 6, 2007 at 7:19 am #

    Please note that the complete ARL Statistics mailing is available at:

  3. Leslie Button December 11, 2007 at 5:55 am #


    Which “categories” of e-books can ARL members count as part of their statistics? Paid/purchased/leased e-books only? Free e-books we have access to (e.g., U.S. Federal documents, National Academy of Science e-books)?

    Thanks for clarifying this!

  4. Carrie Preston December 13, 2007 at 7:26 am #

    I have a question about counting open-access e-serials that are ceased (e.g., old issues are available on the web, but new issues are no longer being published). The webcast and FAQ said that ceased e-serials can be counted “as long as they are being paid for.” Can ceased titles also be counted if they are not being paid for (e.g., they’re open access)?

    The ceased open access titles I am talking about are cataloged.
    carrie preston

  5. Carrie Preston December 13, 2007 at 7:30 am #

    Another question, this one about ebooks: the chat record for the webcast shows a viewer asking whether she could count only ebooks that are paid for in some fashion, ***OR*** whether open-access ebooks can also be counted as long as they are cataloged. Julia Blixrud answered “The short answer is yes,” without making it clear which of the two conflicting alternatives she was saying “Yes” too. Did she mean “Yes, you may ONLY count paid-for ebooks” or “Yes, you may also count free ebooks”?
    thanks again,
    carrie preston

  6. martha kyrillidou December 14, 2007 at 2:41 pm #

    Which “categories” of e-books can ARL members count as part of their statistics? Paid/purchased/leased e-books only? Free e-books we have access to (e.g., U.S. Federal documents, National Academy of Science e-books)?

    All of these can be counted under volumes held as long as they are cataloged.

    Please note that if you are adding one-time large collections of ebooks to your catalog, you should adjust the base figures of volumes in the previous year (we want volumes added gross to reflect the ‘normal’ counts added in the course of the year rather than one time large additions to the colleciton).

    This also answer’s Carrie’s question — ebooks can be counted under volumes held whether purchased or open-access as long as they are cataloged.


  7. martha kyrillidou December 14, 2007 at 2:45 pm #

    Open access serials that are cataloged and currently accessible may be counted even if they are ceased. These e-journals are currently accessible!

    We do want to know how many of those ‘ceased’ ejournals exist — so please provide a footnote. This information will help us evaluate whether we need to revise our recommendation in future years.


  8. martha kyrillidou December 20, 2007 at 1:18 pm #

    Question received:

    1) Since different ISSNs are often assigned to print and electronic versions, why is the ISSN being emphasized as the way to dedupe subscriptions? Do we have a sense of how often this is the case, as opposed to the same ISSN being used for print and electronic?

    Here are my thoughts:

    Implementing deduplication based on ISSN is practically doable and useful — we believe it is the easiest and most practical way to use for deduping serials.

    The fact that the ISSN standard has implemented the decision to assign different ISSN for print and electronic has many underlying assumptions under it, one of them being that the electronic version is commercially viable and independent from the print. This decision reflected the need to identify the electronic and print products separately reflecting authority control issues that go beyond the ‘title’ concept, i.e. titles may not necessarily indicate a unique and distinct entity whereas ISSN should be more reflective of the uniqueness and distinctiveness, and consequently the authoritative existence, of a serial product.

    Irrespective of the viability of the assumptions that have guided the ISSN standard development, libraries are indeed emphasizing the purchasing of electronic versions over print as driven by user demands. So, in the long run we see the need to report the number of electronic ISSN/unique titles as the primary category for reporting serials.
    Clearly print will not go 100% away but it is displaced by electronic products. To the extent that this will be refected in the purchasing behavior of libraries, all print ISSNs for related electronic ISSNs will not be purchased so the electronic ISSN/titles counts is the category that is gaining importance (please note that we had some interesting discussions and dissenting views from some of our non-academic ARL member libraries like the National Library of Medicine where they would have preferred that we track serials in three categories: electronic only, print plus electronic, and print only).

    Another important aspect to remember is that not all serials have ISSNs assigned to them so there is a co-mingling of ISSN and titles reporting in the way Texas A&M implemented the deduplication process for serials. Hopefully more and more serials will have ISSNs assigned to them since deduping ISSNs is indeed a lot more practical (ISSN are by definition unique numbers) and less likely to reflect the kinds of variations you would find in a text field like the title of a serial.

    Martha Kyrillidou

Leave a Reply