Re: Article downloads lower?

[Stevan Harnad <harnad@ecs.soton.ac.uk>]

Fred, May I make a suggestion? You are giving altogether too much weight
to the IOPP download datum, even as you express your reservations about
it. Your message is: "Who says the reduced downloads were because of
Arxiv? Maybe they happened for some other reason."

But the trouble with that is that it implies that if the download
reductions *were* systematically correlated with (hence caused by) Arxiv
downloads, then that would somehow represent a problem, and would somehow
constitute evidence for something else: But what else?

Of course the availability of a self-archived version will reduce
downloads of the journal version -- *especially* in physics, where there
is a special "Arxiv Advantage" in citations, over and above the Open
Access citation Advantage itself: Even articles in Open Access Journals
(as the Journal of High Energy Physics was in its first 3 years), or
articles in journals to which users may have institutional access, tend to
be preferentially consulted online via Arxiv, just because so many other
articles are consulted via Arxiv.

    http://openaccess.eprints.org/index.php?/archives/29-guid.html

What certainly does not follow from that is what ALPSP have been trying to
argue, which is that this somehow constitutes evidence of present or
future *cancellations* of IOPP journals.

Please see Tim Brody's posting today on AmSci about the simple and obvious
solution of pooling repository and publisher stats so as to properly
credit all downloads to the journal in which the article was published,
using OpenURL, so librarians don't come to the wrong conclusion from the
incomplete publisher website download counts that the journal is no longer
of use!

No special pleading is needed to explain away or apologize for the reduce
IOPP website downloads. They are most probably due to increased use of the
Arxiv version. And IOPP have found the right remedy: They are mirroring
Arxiv at the IOPP website. That way they can already start claiming credit
for some of the many access-paths by which their content is being actively
used (since Arxiv has about 15 mirrors!).

Pooled stats will not be long coming too.

Cheers, Stevan

PS The reason IOPP are registering reduced download counts, 14 years down
the road, rather than when Arxiv first started, is because download
counting is more recent, and getting ever more accurate.

---------- Forwarded message ----------
Date: Fri,  7 Oct 2005 19:13:05 EDT
From: "FrederickFriend" <ucylfjf@ucl.ac.uk>
Reply-To: liblicense-l@lists.yale.edu
To: liblicense-l@lists.yale.edu
Subject: Article downloads lower?

It would assist us all in understanding the effect of free repository
content if publishers could make available their download statistics for
articles for which a free repository copy exists. It may then be possible
to compare the publisher statistics with the download statistics for the
equivalent repository copy. I have no reason to doubt IOPP's statement
(from Ken Lillywhite to Lis-e-journals on 5 September)  that "article
downloads from our site are significantly lower for those journals whose
content is substantially replicated in the arXiv repository than for those
which are not", but with all due respect to IOPP the lower downloads could
be due to factors other than the existence of an arXiv copy.

Clearly download statistics do vary between journals even when no
repository copy exists, and the time-line for downloads is also a
significant factor. If public positions are to be based upon such
statistics, as when publishers base resistance to repository deposit upon
lower downloads from their own sites, we need to be sure that the
statistical base is firm. The point is often made that we need more
evidence of the effects of changes in scholarly communication, and JISC
and other organizations have been gathering evidence. Making download
statistics available for examination alongside download statistics from
repositories would be one way in which publishers could work with the
academic community in gathering evidence about the effect of repositories.

Frederick J. Friend
JISC Consultant
OSI Open Access Advocate
Honorary Director Scholarly Communication UCL
E-mail ucylfjf@ucl.ac.uk