Re: Leading academics back UK Research Councils on self-archiving

From: Stevan Harnad <>
Date: Tue, 23 Aug 2005 22:20:30 +0100

On Tue, 23 Aug 2005 Jean-Claude =?ISO-8859-1?Q?Gu=E9don? wrote:

> one may puzzle as to why, when so many articles are already available
> in open access, the rate of use is so low.

It is not clear what it was in Michael Kurtz's posting (to which the
above posting is a reply) that gave Jean-Claude the impression that the
rate of use of the self-archived versions of Astrophysics articles is
low! Michael's point was not that the usage of self-archived drafts is
low, but that the self-archiving co-exists peacefully with the published
journal version. If Jean-Claude thinks the usage is low, he should just
look at the weekly download statistics of, say, the US mirror of arxiv:

Better still, look at Tim Brody's download data for Astrophysics in citebase,
ranking hits by the number of times they were downloaded (for the UK mirror of
arxiv alone)

or in Tim's download/citation correlator/predictor, which will show a good-sized
correlation between downloads and citations in Astrophysics, again despite the
fact that the correlations are based only on UK downloads rather than including US
downloads (as they should -- and I hope soon will, thanks again to Michael Kurtz):

    Brody, T., Harnad, S. and Carr, L. (2005) Earlier Web Usage Statistics
    as Predictors of Later Citation Impact. Journal of the American
    Association for Information Science and Technology (JASIST).

And most important of all, see Tim's data on the 114% citation impact
advantage for those articles that have been self-archived in Astronomy
& Astrophysics compared to those that have not (check "Astronomy &
Astrophysics" and then click "SHOW"):

It is that impact advantage -- which has now been replicated across all
disciplines tested to date, by Chawki Hajjem

-- that is the most fundamental indication of the fact that
self-archiving increases usage, for the only difference involved is
whether the articles (compared within the same journal and year) do or do
not have a self-archived version, freely accessible to all would-be users.

The "rate of use is so low"? What is Jean-Claude thinking?

There is one interesting anomaly, however, unique to Astrophysics, and it has been
discussed before in this Forum:

Astro is unique in being:

        (1) a relatively circumscribed community worldwide, mostly
        located only at the well-funded universities (perhaps because
        of the resource-intensiveness of the research?),

        (2) having a relatively small, "closed" literature, involving a
        small specific set of journals in which all the relevant papers
        and citations appear. As a consequence

        (3) virtually all astrophysicists have institutional site-licensed
        access (i.e., for-fee) to the entire astrophysics literature,
        as confirmed by Tim Brody's finding that (unlike all other fields
        of physics, and unlike all other disciplines)

        (4) astrophysicists self-archive only their unrefereed preprints,
       to get them out early (like other physicists); they do not bother
       to self-archive their postprints, knowing that they will all be
       available to everyone for-fee (as Peter Boyce of ASA has been
       telling me for years).

So the fact that there is *still* an OA citation advantage in Astro, despite all
these factors conspiring against it, is of especial interest.

Michael Kurtz has published a good deal on this topic too:

    Kurtz, M. J., Eichhorn, G., Accomazzi, A., Grant, C.,
    Demleitner, M., Murray, S. S., Martimbeau, N. and Elwell,
    B. (2003b) The NASA Astrophysics Data System: Sociology,
    Bibliometrics, and Impact Author eprint, March 2003, Journal
    of the American Society for Information Science and Technology

    Kurtz, M. J., Eichhorn, G., Accomazzi, A., Grant, C. S., Demleitner,
    M., Murray, S. S. (2004b) The Effect of Use and Access on Citations
    Author eprint, September 2004, Information Processing and Management

> One may also wonder, in the light of [Kurtz's Astrophysics] data
> whether mandating authors to deposit their articles in OA depositories
> would be enough to ensure the success of open access.

One can only wonder what Jean-Claude is thinking, and why! All the
evidence is that even in Astro there is both a usage and a citation
advantage for self-archived papers, just as there are in every other
field. What is true is that it is not Astro (with its 100% de-facto
OA owing to the special factors listed above) that urgently needs the
self-archiving mandate, but all the other disciplines!

> [Kurtz's Astrophysics' figures show that even when depositories (or
> even, in this case, just one depository) contain a very significant
> majority of papers in a particular field, this is still not sufficient
> to encourage actual use.

They show nothing of the sort. They show that self-archived papers still get
downloaded and cited more, even in Astro!

> if such figures were to be multiplied across a significant number
> of disciplines, one could begin to ask whether all the efforts for
> OA repositories are worth it given the benign neglect they seem to
> generate. Supplement to what end, if use remains low?

There is no benign neglect in terms of usage or citation: The benign
neglect is in the failure to do the self-archiving -- and that's what
the mandate is the remedy for. (Please note that once the mandate is
100% successful, there will no longer be a self-archiving advantage,
because all articles will be self-archived. Not even Astro is there yet.)

Supplement to what end? To start enjoying the daily, weekly, monthly,
yearly usage and impact that is still being needlessly lost toady and
has been needlessly (and cumulatively) lost for years:

> if we limit ourselves to OA, OAI-PMH compliant, depositories, and
> with mandating, we will still not be where we want to be. If OA... --
> i.e. as complement and strictly nothing more - is not being used
> very widely, then what is the point? We do want actual use, don't we?

(1) Where we want to be is with 100% of research articles accessible
to 100% of their would-be users, webwide. With 100% self-archiving,
we would be there. And that is the point.

(2) It is only Jean-Claude who thinks self-archived content is not being
used very widely -- in particular, much more widely than non-self-archived
content -- and I think I have already diagnosed why: Because Jean-Claude
is in the thrall of a *theory* -- a theory not about how research can
be made 100% accessible to 100% of its would-be users webwide, but a
theory about how to "reform" the publishing system. This is what makes
him blind to the true benefits (for researchers and research!) of OA.
Jean-Claude is not really interested in OA -- only in OA publishing.

For a critique of Jean-Claude's theory, please see:

    Harnad, S. (2005) Fast-Forward on the Green Road to Open Access:
    The Case Against Mixing Up Green and Gold. Ariadne 43.

Stevan Harnad
Received on Tue Aug 23 2005 - 22:27:55 BST

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:47:58 GMT