Re: Peer Review Reform Hypothesis-Testing

From: Stevan Harnad <>
Date: Wed, 11 Aug 2004 23:18:44 +0100

On Tue, 11 Aug 2004, Heather Morrison wrote:

> While I fully agree with Stevan that we need to proceed with OA given what
> we have immediately, and not wait for all the nuances to be perfected,
> I would like to second Ted's suggestion that "full green" be reserved
> for publishers who provide the final pdf version of an article for OA
> archiving and Open Access, according to the BOAI definition:

The BOAI definition below makes no reference at all to the publisher's
PDF! Indeed, it makes no reference to *form* but only to *content*. The
"full text" means all of the text, figures and tables. It does not mean
only or particularly the publisher's page-images:

> By "open access" to this literature, we mean its free availability
> on the public internet, permitting any users to read, download,
> copy, distribute, print, search, or link to the full texts of these
> articles, crawl them for indexing, pass them as data to software,
> or use them for any other lawful purpose, without financial, legal,
> or technical barriers other than those inseparable from gaining access
> to the internet itself.
> While intermediary steps, as Stevan points out, are most welcome and a
> huge improvement over not having access, it is important to recognize
> that the final, peer-reviewed version is the ideal.

I agree completely with Heather the "the final, peer-reviewed version is the
ideal." Indeed it is the specific target of OA!

But the final, peer-reviewed version -- the postprint -- is exactly what
"full green" refers to in the Romeo Directory right now; and that's not
the publisher's PDF!

> If there is no problem at all with having people access preprints rather
> than final published versions, then perhaps the practices of peer review
> and editing need to be rethought.

Heather, just a moment ago you inadvertently mixed up (1) the publisher's
PDF version with (2) the author's final, peer-reviewed version (the
postprint), which is the version that includes all the revisions from
the refereeing that has accordingly been accepted for publication.

Here you appear to be mixing up (3) the unrefereed preprint and (4)
the unrefereed preprint *plus* an appended list of all the corrections
resulting from the peer review, i.e., the "preprint plus corrigenda"

The "preprint plus corrigenda" strategy is only a logical
reductio-ad-absurdum for that small and shrinking portion of journals that
have not yet given their green light for postprint self-archiving. The
remedy (for this shrinking minority of papers) is to append a link
to a full list of all the corrections mandated by peer review, not to
abandon peer-review!

The purpose of OA is to free the peer-reviewed literature from access-tolls,
not to free it from peer review and its corrections!

And just as the (i) peer-reviewed journal pricing/affordability problem and the
(ii) peer-reviewed journal-article access/impact problem are not the same
problem, so the problem of (a) reforming access to peer-reviewed
journal-articles is not at all the same as the problem of (b) reforming
peer review.

Indeed, it is not at all clear that peer-review needs any reforming! But
even if it does, surely that should come on the basis of actual empirical
studies of ways of improving peer review, and only if and when those
studies actually yield positive results demonstrating that they have
indeed found a better way -- rather than on the basis of untested ideas
that come from trying to solve a completely different problem: the problem
of access to the peer-reviewed literature we have today, such as it is!

    Harnad, S. (1998) The invisible hand of peer review. Nature
    [online] (c. 5 Nov. 1998)

> A great deal of time and money go into these practices; if the OA versions
> are used more, and many do reflect the reviewing and editing - why bother?

I'm afraid I can't find an interpretation of the above passage that I
can understand! Is it about the OA unrefereed preprint or about the OA
postprint (the final peer-reviewed version)? The latter contains the
corrections from the peer review, the former does not. It is the latter
-- the peer-reviewed postprint -- that is the specific target of the OA
movement and the BOAI definition cited above, not the former (the
unrefereed preprint).

Yes, time and money do go into implementing peer review, and that is an
essential service if we are to have the peer-reviewed literature to which
we are here talking about providing Open Access. No one has proposed
doing away with the peer review; we are talking about self-archiving
the postprint so as to provide OA.

(The fact that in disciplines that also self-archive unrefereed preprints
those unrefereed preprints also get a lot of usage is a bonus, but it is not
what OA is about. OA is not about jettisoning peer review and having only
preprints. It is about self-archiving postprints: Not necessarily PDFs, just

> There could be arguments that there are more efficient means of peer
> review in this day and age.

Before we even consider abandoning the peer-review system that has
given us the 24,000 peer-reviewed journals (to whose contents we are
here trying to provide OA) and instead trying another, untested system,
surely what we need is not *arguments* that that new system would work,
and would give us a research literature of at least equal quality,
but *evidence* that it will actually do so.

But, until further notice, are we not discussing the problem of
access/impact here, rather than the problem of quality-control? Where is
the evidence that there is a pressing quality-control problem to remedy?

Or is it price (the price of quality control)? But we are talking here
about the green road to OA, author-instition self-archiving, in which
the cost of quality-control continues to be paid the old way -- through
access-tolls (subscriptions, licenses) paid by those institutions that
can afford to pay them, as now. We are talking about *supplementing*
that system with OA, for all those would-be users whose institutions
who cannot afford the access-tolls -- with that OA being provided by
authors, self-archiving their own papers, published in those peer-reviewed
toll-access journals!

So the cost of peer-review does not even come into it. Cost-cutting is
for the other road to OA, the golden road of OA journal publishing. But
even there, it is classical peer review that is being used, not any of
the hypothetical variants that Heather mentions below:

> Send a message to this listserv, for example,
> and there seems to be a very high probability that it will be thoroughly
> and publicly reviewed, with counter-arguments often quickly provided
> by other participants. Could having everyone working in a reasonably
> narrowly defined area share in a listserv, with editorial commenting
> duties revolving perhaps, with research results shared first through the
> list, then posted to a final archive after confirmation by an authority
> selected by this group of colleagues that any necessary changes have been
> made? Couldn't a system like this be an extremely cost-efficient way to
> make use of the fact that the great majority of this work is done on a
> voluntary basis?

Heather, many have raised this question. No one has bothered to test it to
find out what the answer is. Let someone go and do that, with a sufficiently
large, representative and scalable sample of the peer-reviewed literature,
and for long enough to draw reliable conclusions -- and then come back and
tell us what they found!

Till then, are we not discussing how to reform *access* to peer-reviewed
research rather than hypotheses about how to reform peer review?

> As for authors making their own corrections, the results (quality of
> what people are reading) is unknown at this point in time. We don't
> know whether authors will actually make the corrections, how accurate
> they will be (possible sources of error ranging from simple copying
> problems to disagreement with reviewers and editors and refusal to make
> suggested changes), or whether authors might decide, in the course of
> making corrections, to include additional changes such as updating -
> which might provide new, useful information, but could also add sources
> of error, if the author is adding information that really should have
> been submitted for peer review.

You are now talking about how scholarly practises will evolve in an
OA world. No doubt they will evolve. But right now they are not the
issue, as we are as yet far from that OA world. I think we should be
doing everything to get us there as soon as possible, rather than trying
to second-guess how our scholarliness will change once we get there.

When what we *should* be doing (but are not doing!) is self-archiving our
postprints (or our preprints + corrigenda) as soon as possible, surely it
is not helpful to pause instead to contemplate whether those postprints
(which we are not yet self-archiving!) would be good enough! (Maybe
we should wait till we can self-archive the PDFs?) Or whether those
postprints (which we are not yet self-archiving!) are even necessary at all!
(Maybe we should abandon peer review, and post our preprints to lists

Postprints (to repeat the BOAI definition of OA) are the target. (And
the preprints-plus-corrigenda strategy is a makeshift solution to cover
the temporary exceptions, if any, for now.)

> These kinds of issues are likely highly
> variable from discipline to discipline - physics does seem to be doing
> very well with OA preprints, for example, whereas in some areas of the
> social sciences where there is tremendous potential for bias on the part
> of a researcher, peer review is probably much more important.

All disciplines currently practise classical peer review, physics included.
The only thing that has changed -- anywhere -- is access.

Physicists self-archive pre-refereeing preprints *in addition*
to submitting them for peer review, and they self-archive the
post-refereeing postprints too (and sometimes intermediate drafts, as
well as postpublication corrections and updates). No discipline today
has an alternative to peer review. And (more important and relevant), no
discipline today would fail to benefit (in research impact, productivity
and progress) from providing Open Access to 100% of its peer-reviewed journal

Stevan Harnad
Received on Wed Aug 11 2004 - 23:18:44 BST

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:47:33 GMT