Monday, October 16, 2006

The use of usage

I came across an interesting article by John Ewing, Executive Director and Publisher of the American Mathematical Society, published in the Notices of the AMS this month (October issue). The article is entitled "Measuring Journals" and discusses impact factors and usage statistics.

Statistics are funny things. The decline of the birth rate in Western Europe coincides with the decline of the stork. Imagine the possible conclusions if you don’t understand the statistics (though they may confirm long-held beliefs). Usage statistics have to be understood before they can be used to come to any meaningful conclusions, if ever.

Even if we do understand user statistics sufficiently, how reliable can they really be? Not very, is Ewing’s conclusion. We have to be extremely careful when we use such ‘objective’ quantitative data for qualitative conclusions. Ewing further says that “Distrust of ‘subjective’ scholarly judgment is a modern disease – one that is profoundly anti-intellectual.” I would add that blind trust in ‘objective’ measurements is equally profoundly anti-intellectual.

Suppose we can be confident that we understand the statistics, does usage determine the value of journals and articles in the first place? I’m aware of the adage publish or perish, but not of one that says read or rot or download or be damned. Isn’t the value therefore more in the availability of a publication than in its usage? Isn’t there a strong value element of ‘just-in-case’ in scientific literature (like the value of insurance – where you’d probably avoid actual ‘usage’)? Isn’t there a strong value element in just making sure that research results are properly recorded (like the minutes of important meetings – they are not often read a lot, but it’s crucial that they are made)? The ‘minutes of science’ as I used to call it in the mid-nineties?

Isn’t it so that a manuscript with potentially interesting information is only made actually interesting if the outcome of a process of peer-review shows that it’s been formally accepted and acknowledged by the scientific community as worth adding to the body of literature, and labelled as such (with a journal imprimatur)? And isn’t there then more value in the label it carries (imprimatur, certification, however one calls it) than in the information itself (which may well already be out there in cyberspace and often is)? And isn’t that mainly a value for authors (remember: publish or perish) and their careers and future funding prospects rather than for readers (remember: there’s no read or rot)?

As an information exchange, many journals may already have lost their role. The internet is definitely taking over. But ‘usage’ of a journal as a formal recording and validation service has not disappeared. Arguably, that service is more valuable now than ever, given the difficulty of establishing the integrity of information available on the web.

In my view that means that the economic underpinning of journals by placing a monetary value solely on download usage is outdated. Much of the monetary value should, instead, be placed on the service of formally publishing the material. In an ‘author-side-payment’ model that is explicitly the case and such a publishing model also means that open access, i.e. universal availability, can be the natural condition of the formal, officially published articles.

Jan Velterop

1 comment:

  1. hi Jan -

    In my view, relying on usage can be more than outdated, it can be downright dangerous to scholarship when used as a factor in pricing, or subscriptions / cancellations. Usage based pricing inevitably discourages use; there is no good reason to discourage scholarly reading in the electronic environment. Based subscription or cancellation decisions on usage could be very detrimental to smaller research communities - including those who research endangered species and rare diseases. Small losses of small research communities can add up to a huge loss in diversity.

    My thoughts on this topic can be found in my book chapter, "The implications of usage statistics as an economic factor in scholarly communications", in "Usage Statistics of E- Serials" by David Fowler (editor). Published by the Haworth Press [2005]. Part of the "Haworth Series on Serials Librarianship and Continuing Resources", available in the SFU D-Space at: Or, in E-LIS at:

    If anyone is considering using usage statistics for pricing / purchasing decisions, please read this chapter - not for my personal impact, rather for what this model means for scholarship and hence the world.

    Fortunately there is no need whatsoever to think about usage-based models. We have open access.