RE: NEJM editorial on open access

[John Sack <sack@stanford.edu>]

Coincidentally, I do have some recent analysis on the question you ask.  
Thanks for asking a question I have an answer to!

About 200 journals that HighWire works with participate in the "Free Back
Issues" program that began in 1997 with JBC, JCB and PNAS deciding
together that their research publications could benefit non-research
programs by making content free after some period.  There are now over
770,000 articles made free by that program at HighWire.

The information about what journals participate, and after what "delay
period" is available here:  
http://highwire.stanford.edu/lists/freeart.dtl (This page also lists all
the free trials, so it is a looonnngggg page.)  I include this page not
only to publicize the program, but so others can re-do the analysis I
mention in the next paragraph, which is the answer to your question:

Most (85% as I recall) of those 200 journals make their content free after
a year (a few after 18 or 24 months, sometimes reviews journals have
longer delays).  Included in that, 23, or 12%, of the 200 journals that
make content free after a delay period make it free in six months or less.  
(I am excluding from the percentage the sites that are totally free, since
a number of them are not original-research journals; but there are 24 such
sites if you want to add them back in; that would make it 23% are free or
6 months or less.)

I don't believe that one can conclude that there is evidence that six
months works for all (or even many) journals, based on what we can see
now.  One also can't conclude that it does not work (except anecdotally as
noted below); it obviously does work for some.  I do hear a number of
publishers wondering why someone would think one size would fit all, any
more than one subscription price (or one OA fee...:) would fit all
journals.  On the other hand, I think one could reasonably say the data
shows that 12 months has been successfully adopted by most *participating*
journals (participants are only a portion of the journals we work with; a
significant number of journals make no content free after a delay period).

I do recall one major journal that had a delay period of two months found
that there were enough library cancellations that they moved the delay to
six months.  However, another journal that started at two months has
stayed there and happily so. The former journal heard from librarians that
they didn't understand why a library should pay for a twelve-month
subscription when the general public got the journal free after two
months; and the cancellations showed this was the logic being applied.  
Over the years several journals have raised their delay periods, and
several have lowered them; because HighWire isn't involved in the business
but only the technology of these decisions I can't say why.  But if the
delay period is in part based on librarian attitudes (and budgets) then
this too will vary over time.

Two years ago, I was talking with an Australian librarian about a
six-month delay period, and his comment was "why, six months is just a
slow boat!".  He indicated that they would likely selectively cancel
journals that were free after six months (presumably applying some
differentiation such as you describe, but he didn't say what).

I once had a theory that perhaps journals that published weekly or twice
monthly might find that six months worked for them; and journals that
published monthly might find that 12 months worked for them.  But the data
shows no correlation between frequency of publication and delay period.  
Since the data doesn't cooperate, I've gone back to tea leaves.  There
might be a correlation with impact factor, or with ISI 'immediacy' factor.

Most of the 200 journals provide the free content on their journal site,
and nowhere else.  But some 30 are also in PubMed Central; about half of
these use the PMC "host" model in which the content is accessed both at
PMC and at HW, and about half use the (no longer offered) PMC "link" model
in which the content is indexed at PMC but viewed at the publisher's HW
site).  I mention this seemingly obvious fact because many publishers
believe they gain new readers for their content by attracting
readers/attention to their site through offering free back content, and
this is significant to many publishers in the free back issues program.  
The NIH plan, as they read it, does not offer this 'read the content on
the journal site' model (it offers publishers who deposit final, published
copy the option of a link, but the article can still be viewed on the NIH
site).  And to bring my long email and this whole thread full circle, that
is what the NEJM editorial is suggesting: a link from the NIH repository
would deliver the article from the NEJM website for publishers who prefer
that approach.

HighWire supports many different publishers, and I wouldn't speak for
them, especially in their variety.  But I hope the data and interpretation
above is useful in understanding why there is such a variety of opinion
about the NIH plan.

Sorry for the long note in reply to a simple question.

John Sack, Director
        HighWire Press, Stanford University
        Phone: 650-723-0192; fax: 650-725-9335
        http://highwire.stanford.edu/~sack
        sack@stanford.edu
        
PS: Here's an interesting factoid: just about half of the NIH-funded 
research studies in 2003 were published in HighWire-hosted journals.  And 
of that half, 85% is free within a year.  So the NIH plan is at least 
partly accomplished already in spirit in the sense that the literature is 
freely accessible, and of course it is indexed in PubMed and searchable in 
google as well.  But it isn't free within 6 months, and it isn't mostly in 
PMC (though entirely in PubMed).

####

On Thursday, October 21, 2004 8:50 PM -0400 David Goodman 
<David.Goodman@liu.edu> wrote:

> Dear Joe and jcg,
>
> It is not correct to say the the free public available of an uncopyedited
> version after 6 months will cause major cancellations.  I hope someone
> from Highwire will correct me if I am wrong, but it is my understanding
> that for the many journals they distribute, free availability of the full
> published version after 6 months has not lead to significant subscription
> losses for the journals concerned.  I think that this can be demonstrated
> even without quantitative data: almost all the publishers have continued
> this availability, and they surely would not have done so if it were
> damaging to their subscription list.  The length of this period for these
> journals is not compelled, but is their free choice.
>
> This is not unreasonable for first rate journals, such as those. Those who
> want them most want them immediately. It might be significant for journals
> which nobody really reads much anyway: I would say that in most scientific
> subjects, if researchers do not care whether they see a journal's articles
> within the first few months, the journal is essentiallymoribund, and will
> be a prime target for cancellation in any case, regardless of OA.  There
> should be no regret it if OA causes a faster cancellation of the titles
> that are both expensive and virtually useless-- even their publishers may
> be better off without them.
>
> But at least in biomedicine, it has been proven that 6 months offers
> sufficient protection for a good journal--what remains to be proven is
> whether in biomedicine even shorter periods are not sufficient.  In
> physics, it has been proven that no embargo period at all is necessary.
>
> I think OA is possible for all publishers of good journals, commercial or
> non-commercial, and I would not discourage any. Recent years have seen the
> massive flight of capital into , not out of, the commercial publishing
> industry, as the purchases of Kluwer and Springer have demonstrated, These
> purchases were made with the full awareness of the forthcoming possibility
> of OA.  I even suggest that it is possible that "Gold" OA Journals may
> prove particularly suited to the commercial publishers, because they have
> the capital to keep them going over the transition, whereas not all
> societies do.
>
> There are a wide range of possibilities with the introduction of OA, and
> we shall soon see which of them will be realised. But there are some that
> are not probable, and the immediate massive cancellation of good journals
> is among them--I would regard it as about equally likely as the immediate
> adoption of OA Journals by all publishers. There may be fundamental
> changes, and they may happen during the next 5 years or so, but this gives
> publishers of all types sufficient time to adapt the way they do
> business.   I cannot imagine a publisher so unaware that it has not begun
> planning some time ago.
>
> Dr. David Goodman
> Associate Professor
> Palmer School of Library and Information Science
> Long Island University
> dgoodman@liu.edu