Re: The True Cost of the Essentials (Implementing Peer Review)

From: Stevan Harnad <>
Date: Wed, 25 Jul 2001 22:18:33 +0100

On Wed, 25 Jul 2001, Arthur Smith wrote:

> most of the cost is in rejecting articles that don't fit the journal's
> criteria for importance and general interest, not merely the criterion
> of correctness or "quality".

I think that (almost by definition) the higher the rejection rate of a
journal, the larger the proportion of the overall peer review
implementation costs spent on processing ultimately rejected
manuscripts (and rejection rates as well as impact factors tend to
correlate with journal quality, although it would be nice to see hard
data on that, across disciplines). Rejection costs are hence a part of
the true total cost of the essentials (whether that total proves to be
10% or 30% of current S/L/P revenues).

It will require some experimentation if ever the S/L/P revenue from
selling the add-ons (paper edition, publisher's PDF, online
enhancements) shrinks to where it no longer covers the cost of the
essentials (peer review). One way to recover the rejection costs would
be to wrap them (like shop-lifting costs) into the costs per accepted
paper; another way would be to charge a submission fee that can be
credited to the acceptance costs should the paper ultimately be accepted.
But the matter is delicate, because (like peer review modification or
reform) unprecedented and untested.

> perhaps $640-$740 per accepted article.

That still sounds rather high to me (JHEP, with only ONE journal,
reports $500: surely there should be some economies of scale here) but
let's provisionally accept that figure. What is the current total
revenue per article (piecing together all collective inputs from
S/L/P)? I assume you are saying it's three times that. Fine. That
sounds like more than enough to cover the essentials if the market for
the add-ons should ever dry up because of self-archiving.

> sh> Arthur, all I advocate is ubiquitous author/institution self-archiving,
> sh> now. Whether the actual figure for peer review turns out to be 10% or
> sh> 30% does not matter very much.
> Then PLEASE stop repeating the 10% or $200 figure. A factor of 3 or more
> error in your statements diminishes your credibility, and severely
> misleads anybody who believes your number.

Do you think the APS estimate is a better average for the 20,000+
refereed journals and their 2,000,000+ annual articles? (I am not
asking ironically: I really wonder how representative you think the APS
bottom line is. We are talking about averages here, after all, and
S/L/P revenues vary from $500 per article to $4000+ from journal to
journal, and publisher to publisher. Submission and rejection rates as
well as processing demands vary too.)

> > > but if we don't look at other models
> > > we'll all just keep doing what we're doing, and it's always going to
> > > cost about the same, and libraries or institutions or whoever foots the
> > > bill will just have to keep it up. And Open Archives will continue to be
> > > basically irrelevant.
> >
> > Alas I could not follow the logic of that. Self-Archiving can and
> > should free the refereed literature now. How to cover the essential
> > peer review costs if and when S/L/P can no longer do so is a bridge we
> > can cross when we get to it. Even at 30% there would be plenty to cover
> > it. So what are we disagreeing about? And how does it follow that S/L/P
> > continues to have to be paid at current levels, and that self-archiving
> > is irrelevant?
> Either author self-archiving is sufficient to provide access to the
> peer-reviewed literature, or it isn't. If it is, then journal S/L/P as
> it currently exists can be cut by libraries and institutions with no
> loss to those institutions, and publishers will have to go begging to
> keep doing the peer review stuff, or will simply abandon it. If author
> self-archiving is NOT sufficient (which is what I interpreted Sally
> Morris' comments to imply) then my statement follows, and the open
> archiving is basically irrelevant to the journal pricing problem which
> it is supposedly addressing. And looking at the evidence, as
> demonstrated in fields such as High Energy Physics for which even
> commercial journals seem to be still quite viable (after 10 years of
> pretty complete author self-archiving), author self-archiving is not
> sufficient in this sense.

I still can't follow this at all. Are we talking about the eventual
author self-archiving of the entire refereed literature (20K journals,
20M articles annually), and what the eventual impact of THAT might be
on S/L/P revenue? Or are we merely talking about the pitifully small
portion of the annual 20M articles self-archived so far, which is still
only about 50K annually, most of it in physics, amounting to only about
30-40% of the total physics literature and not destined to reach 100%
of that until the year 2011 at the current linear growth rate:

I see the problem as that of awakening researchers to the benefits (in
terms of visibility, accessibility, and hence potential impact) of
freeing access to their research online through self-archiving. In
other words, the problem is getting the 20M up there, along with the

If that doesn't significantly reduce the demand for the S/L/P add-ons,
that's just fine. Everything can continue as now, but the refereed
literature will be accessible online for free.

If it does reduce the S/L/P demand to where the essential costs of peer
review (be they 10% or 30% of current S/L/P expenditure) can no longer
be covered out of S/L/P, then they can be paid for out of the
author/institution's annual windfall S/L/P savings as peer-review costs
for outgoing papers instead of S/L/P costs for incoming papers.

The arithmetic works whether it's 10% or 30%. There will be
institutional windfall S/L/P savings if S/L/P demand drops sufficiently
to make it insufficient to cover peer review costs out of S/L/P. It
does not seem to me to require too much brainwork to devise a system
whereby part of the annual windfall S/L/P savings on incoming papers at
each research institution is redirected toward covering outgoing peer
review costs. Necessity is the Mother of Invention (though I think some
pre-planning could leverage the transformation, perhaps with collective
institutional consortial support, if there are signs that the
transformation is taking place, so as to keep the process smooth and

By the way, I do not consider the Self-Archiving Initiative (which is
much narrower than the Open Archives Initiative) to be motivated by "the
journal pricing problem" (aka "the serials crisis"), although, once
successful, it will solve that too. It is motivated by the goal of freeing
refereed research from the no longer necessary or justifiable Gutenberg
barriers to access and hence impact (S/L/P):

> Author self-archiving is fine in itself, and certainly it can greatly
> increase access for remote and poorer places in the world, but it won't
> be doing anything to cut costs to libraries unless the 3rd party forum
> represented by peer review and journal editorial processes itself starts
> to be absorbed and overlayed in some new and innovative manner.

I again could not follow this. Here is a certainty: Freeing the annual
20M will make it accessible to everyone online, rich or poor (and no
institution, not even Harvard, can afford anywhere near all 20M: ).
That desideratum has face validity.

Whether or not library S/L/P demand will fall significantly once a
significant portion of the 20M becomes accessible free is only a
hypothetical question on which nothing significant hinges now
(though there is no harm in thinking ahead).

Researchers should self-archive all their refereed research now; we can
worry about what to do about the S/L/P bridge if and when it ever shows
signs of heading for collapse. We can prepare some plans now, to be
sure, but not if they deter us from launching the 20M skyward.

Stevan Harnad
Professor of Cognitive Science
Department of Electronics and phone: +44 23-80 592-582
             Computer Science fax: +44 23-80 592-865
University of Southampton
Highfield, Southampton

NOTE: A complete archive of the ongoing discussion of providing free
access to the refereed journal literature online is available at the
American Scientist September Forum (98 & 99 & 00 & 01):

You may join the list at the site above.

Discussion can be posted to:
Received on Wed Jan 03 2001 - 19:17:43 GMT

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:46:11 GMT