(wrong string) £1.5bn a year

From: Peter Banks <pbanks_at_DIABETES.ORG>
Date: Tue, 27 Sep 2005 09:50:34 -0400

It is not a silly question. Dr. Harnad and his colleagues address it in their paper "Citation Impact of Open Access Articles vs. Articles available only through subscription ("Toll-Access")" http://citebase.eprints.org/isi_study/ They do say they have factored out self-citation.

Whether the issue of self-citation or any other aspect of their work is convincing is another matter. Although Peter Suber recently claimed (in a letter to the Washington Times) that "Study after study has shown that free online access increases the impact of research literature, as measured by citations, 50 percent to 250 percent," I am not sure what "study after study" refers to, though is clearly is a reference to Harnad's work.

Dr. Harnad has provided one other refererence (http://www.crsc.uqam.ca/lab/chawki/graphes/EtudeImpact.htm ), so perhaps "study after study" means literally that: two studies. Or maybe there are more, but I can't find references. Neither of the studies above appears to have been peer reviewed or published other than by preprint. Indeed, in the first study, the authors make this disclaimer: "Warning: The data presented here are preliminary unrefereed results that are still being analyzed and corrected (we welcome any suggestions or questions). This is not yet the "definitive" version of our findings. Please do not cite them without consultation with the authors."

I would encourage interested parties to take the authors up on their invitation for (much needed, in my view) peer review.




Peter Banks
Acting Vice President for Publications/Publisher
American Diabetes Association
1701 North Beauregard Street
Alexandria, VA 22311
703/299-2033
FAX 703/683-2890
Email: pbanks_at_diabetes.org

>>> harnad_at_ecs.soton.ac.uk 09/25/05 3:12 PM >>>
On Fri, 23 Sep 2005, Sally Morris (ALPSP) wrote:

> The problem lies with Stevan's 50% figure - apparently picked out of the
> air, and with no factual basis whatsoever - for the increased 'return on
> investment' if research is OA. I don't find it very convincing to base
> such sweeping conclusions on a completely unsupported figure

Picked out of the air? I reported (and provided the references and URLs)
the strong new empirical evidence that open access articles consistently
receive 50%-250% more citations, comparing always within the same journal
and same year. Here are some summary data at the discipline level:

In each case the two percentages will be

    (%OA) the percentage of OA articles among all articles (OA and non-OA)
    in the same journal/year: OA/(OA + nonOA) articles

    (%OAc) the percentage gain in citations for OA article compared to
    non-OA articles in the same journal/year: (OA/nonOA) - 100% citations

 From: http://www.crsc.uqam.ca/lab/chawki/graphes/EtudeImpact.htm

                      %OA %OAc
    Administration 6% +180%
    Economics 14% +49%
    Education 5% +77%
    Psychology 6% +93%
    Management 8% +68%
    Health Science 5% +57%
    Social Science 14% +126%
    Biology 14% +30%
    (other disciplines: data still being gathered, samples still too small)
    
 From: http://citebase.eprints.org/isi_study/

    Astrophysics 24% +114%
    Nuclear/Particle 38% +120%
    (other disciplines: data still being gathered, samples still too small)

I then took a low-end conservative figure for OAc at 50% and applied it to
the conservative figure of 85% not yet self-archived, to yield 50% x 85% x
£3.5.bn = £1.5bn worth of loss of return (in terms of citations) on the
RCUK's £3.5.bn annual investment.

As noted, it is not the number of articles published annually (about
130,000) that represents the return on the UK's research investment; it is
how much those articles are used, applied, and built-upon. Research
published but not used, applied and built-upon is research that may as
well not have been done or funded at all. The citation counts are measures
of the degree to which research is used, applied and built-upon --
"research impact."

The UK is losing 1.5bn worth of potential research impact annually (on our
conservative, low-end estimate) for the 85% of it that it is not yet
self-archiving (another conservative estimate). The RCUK open-access
self-archiving mandate -- *if* it is not hobbled into an open-ended
embargoed-access policy, as the NIH policy proposal was -- will remedy all
of this needless loss of research impact and return on the UK public
investment in research.

    "Please Don't Copy-Cat Clone NIH-12 Non-OA Policy!"
    http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/4307.html

    "Open Access vs. NIH Back Access and Nature's Back-Sliding"
    http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/4312.html

Please note that I did not say the UK was getting *no* return on its
research investment: Even non-OA articles get used and cited -- but only
by those users whose institutions can afford the toll access to the
journal version. The empirical 50-250% citation-gap corresponds to the
loss of the potential research impact from those users who are currently
denied access. Self-archiving the author's version is done to maximise
usage, impact, and hence the return on the public investment, by making
the research accessible to those access-denied would-be users too.

But just as toll-access is not open access, and fails to maximise research
impact, so embargoed access is not open access and fails to maximise
research impact. The self-archiving must be required to be done
immediately upon acceptance for publication. To allow delays of 3, 6, 12
months or more would simply be to return to the needless loss in the
return on the public investment in research that the RCUK self-archiving
mandate is intended to remedy.

So: No embargo, of any length at all. What can be allowed instead -- with
some loss in efficiency, but no significant loss in impact -- is the
immediate, required self-archiving of the full text and the metadata
(author, title, journal, date, etc.), with the access-setting for the
full-text to "open access" being merely encouraged, but not required. If
an author prefers to set access to the full-text as "institution-internal"
access only, the metadata are still visible and searchable to all, and the
full-text can still be harvested and inverted by google without displaying
it (as google already does with books).

Would-be users can then email the author for an eprint: Somewhat slower
and less efficient than direct click-through access, but good enough. As
over 90% of journals are already green on self-archiving, fewer than 10%
of articles will suffer from this inconvenience, it will be bad press for
the non-green publishers, and the authors will soon tire of doing the
keystrokes to keep emailing the eprint -- and will simply do the last
keystroke, switching access from "institutional" to "open."

    http://romeo.eprints.org/stats.php
    
Stevan Harnad

American Diabetes Association
Cure. Care. Commitment.


Visit us at http://diabetes.org
Or Call 1-800-DIABETES (800-342-2383)
Received on Wed Sep 28 2005 - 02:22:38 BST

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:48:02 GMT