Craig et al.'s review of the OA citation advantage

[Stevan Harnad]

Excellent points by Matt Hodgkinson. A few supporting comments:

On Fri, 25 May 2007 matt.hodgkinson at biomedcentral.com wrote:

> in a famine it is no good if food is in the shops, but the prices are too 
> high for the starving to afford it.

Spot on.

> I don't want to pay $25-50 to read an article I'm not sure is worth the money...
> Indeed, if it is not immediately available online then even a visit to the 
> library... I would avoid if possible...

And it's virtually certain that huge quantities of potential usage and
impact are being lost daily, worldwide, for this very reason. Indeed, a
component in the OA usage/impact advantage is a *competitive* advantage
(CA): The articles that are not yet freely accessible online lose out
to the ones that are. CA is not the only component in the OA advantage,
nor necessarily the biggest one.  And CA (along with the self-selection
Quality Bias QB) will of course vanish completely once everything is
OA. But for now, CA is an extra -- and potentially substantial --
competitive edge that the OA articles have over the non-OA ones while
much of research is still non-OA.

> I don't quite understand something about Early View - is it solely an 
> effect of preprints / self-archiving? 

No, it definitely applies to all OA papers, whether preprint or
postprint, self-archived or in an OA journal. Early Access means
having the OA advantage earlier. The earliest possible moment for the
refereed draft is the moment when the final version is accepted for
publication; that is the *latest* time at which it should be made
OA. (Until then there may still be changes and corrections from the
refereeing; and in many fields cautious users will not want to risk
relying on the unrefereed preprint. So preprint self-archiving must
be discretionary; it is postprint self-archiving that must be mandatory.)

I strongly doubt the claim that Early Access just means phase-shifting
the lifetime citation expectancy of an article, i.e., that it's the same
number of total citations, but they just start happening earlier. I
think it might look like that for fields that are virtually 100% OA
already, like astrophysics: there it has been reported that the citation
curves look the same for articles that are and are not self-archived
as preprints, just that for the preprinted ones the curve starts
earlier. 

What this leaves out is when the curve *ends*! Two wave-fronts may look
the same, apart from a phase difference, but then there's the question of
the long-term total area under the wave.  The way paper uploads generate
downloads -- which then generate citations, which then generate more
downloads, which generate more citations, etc. -- suggests that this
interactive cycle increases not just the onset time of citations but
the total area (citations) under the curve. Nor does it stop there:
Other research is going on in parallel. If it is obvious that it is not
irrelevant to the usage and impact of a finding whether it is published
two months before it is needed for a related study by another researcher,
or ten years after, then it should not take much imagination (just a
change in time-scale) to see how Early Access does not just mean earlier
citations but more citations, because of the widening self-potentiating
cycle of research.

And this of course applies to both preprints and postprints: An article
that is published at time T but only made OA at time T + 12 months
(embargo) stands to lose a good deal of its potential impact (especially
in fast-moving fields) -- some of it lost forever; and meanwhile research
loses potential widening cycles of progress.

> Early View appears to be a somewhat complicated way of saying 
> that if an article is available earlier, it can be read and cited sooner. 

Which is in turn a somewhat complicated way of saying that if an article
is accessible, it can be read and cited, and the more it is accessible --
whether more widely or earlier -- the more it can be read and cited. In
other words, OA applies to both time and space: The sooner and the more
widely findings are accessible, the sooner and the more widely they can
be taken up, applied, built upon, used, and cited. Early Access benefits
are merely a particular case of OA benefits.

It is only to those who are straining to make us swallow publisher embargoes
-- as if they made no difference at all to research usage, uptake,
impact, and progress -- that these banal truths will be anything less
than obvious.

> Is the rapid dissemination of science not a good thing, and should this 
> result not encourage all authors to deposit preprints and postprints?

Of course it is, and should. The only ones who would have us think
otherwise are those who feel their revenues might be put at risk by such
deposits (which, eventually, they indeed might).  But instead of just
coming out and saying that -- "Please don't self-archive, because it
might make me lose some subscription revenue" -- they try to persuade
researchers not to self-archive because it wouldn't make any difference
to research.

This strategy calls to mind nothing less than the efforts of polluting
industries to persuade the public that the pollution makes no difference
to their climate, or the efforts of tobacco companies to persuade
smokers that the smoking will make no difference to their health. The
strategy is essentially the same as that of OJ Simpson's Dream Team:
Simply take every piece of empirical evidence (that is unfavourable to
your client), find some ad hoc flaw in it, no matter how trivial, and
crank and spin that so as to sow a seed of doubt in every instance. Such
a strategy worked for the tobacco industry until the evidence became
overwhelming. But meanwhile, smokers needlessly lost years of health,
just as research is now needlessly losing years of impact and progress.

I have for years been restraining myself from making these analogies
with the tobacco and pollution industries, because it seemed too shrill:
impact, after all, is not as important as health. Maybe, maybe not. But
what is making me less inclined to continue to be so restrained and
charitable is the relentless (and successful) lobbying by the publishing
industry against Green OA mandates. The motivation is identical: Do and
say whatever it takes to protect your revenue streams, whether it's at
the cost of research impact or health impact.

The gloves are now off...

Stevan Harnad

PS I don't mean Sally, of course, but the publishing industry's pit-bulls,
who have so far successfully lobbied the DTI in the UK, NIH in the
US, the Bundesrat in Germany, the EC in Brussels and the Industry and
Finance ministries in Canada. OA has no lobby, but it has a far, far bigger
constituency, which needs merely to be rallied to show its collective strength:
researchers, research institutions, research funders, the vast R&D industry,
and the public whose taxes support the research.

    http://www.nature.com/news/2007/070122/full/445347a.html
    http://www.ec-petition.eu/index.php?p=index
    http://www.publicaccesstoresearch.org/index.html

> > From: Sally Morris (Morris Associates) 23 May 2007 14:58
> > To: SPARC Open Access Forum
> > Subject: Re: Recent research tempers citation advantage of open access
> > 
> > I don't follow this at all.  EV is to do with the article being available
> > sooner, not more widely (that would be the 'OA advantage', if any).
> > Articles in OA journals are available no sooner than those in conventional
> > journals - i.e. on publication.