What about delayed open access

[Stevan Harnad]

It has been a shortcoming of most studies on both the extent of OA and
the size and extent of the OA impact advantage that they take into
account the date of publication of both the cited articles and the
citing article, but not the date on which (Green) OA articles were
made OA.

Finer-grained studies are underway. I think anyone with an open mind
will agree that if OA is  beneficial to research impact and progress
(i.e., if denying access to non-paying users is deleterious to overall
research impact and progress) then OA is desirable immediately upon
acceptance for publication, and not only after an access-denial
interval (embargo).

There are already at least two kinds of analyses that show this: (1)
The studies of Kurtz and co-workers on the Early Access advantage for
prepublication preprints in astronomy and physics. (Earlier
self-archiving does not just generate downloads and citations earlier,
but that earlier usage and citation fans out into more downloads and
citations overall, accelerating research not just by reaching a fixed
"impact quota" sooner, but by increasing its impact quota.) (2)
Brody's studies have shown that not only does self-archiving lead to
earlier citations, but that higher rates of downloads early on are
correlated with higher rates of citations later.
http://eprints.soton.ac.uk/40867/02/chapter5distro.pdf

Some comments below:

Bo-Christer Bjork wrote:

> In the current OA barometer project we're now in the final stages of our
> empirical work trying to establish what part of the 2008 peer reviewed
> article production is available as OA. Overall it seems the share available
> in journals and as e-copies is around equally big. What is particularly
> interesting is the split into different types of channels also inside gold
> and green. We will publish the results in due course but I would already now
> point out that we have found a perhaps surprisingly large amount of articles
> which have become OA on toll-gate publishers sites after a delay of 12
> months.

It is crucial, then, to classify these "delayed OA" articles made OA
by their publishers after an embargo period as Gold OA, not as Green
OA. They are only Green OA if self-archived by their authors.

Both the Green OA and Gold OA tallies should be classified by posting
date, otherwise we will get a misleading impression of both the OA
proportion per year of publication and the annual growth rate for OA
self-archiving. [Note that with publishers making their own articles
OA after an embargo, this means *both* Green OA and Gold OA (and not
just Green OA) need to be analyzed in terms of deposit date and not
just publication date.]

> Very often you can only find this out after trying out with more
> recent articles, since the publishers in question don't seem to advertise
> the delayed OA. It becomes particularly intriguing when the same publishers
> also practice "Open choice" for individual articles. Why pay if all articles
> become free after 12 months anyway?

Better question: Why pay for immediate Gold OA, or wait for delayed
Gold OA, when you can self-archive and provide immediate Green OA?

> I think we should take note of this and accept delayed OA as a viable form
> of Open Access. What is in fact the difference between this and a repository
> copy posted after an embargo of 12 months.

A big difference. See the Kurtz and Brody studies for a foretaste. (Of
course the degree to which usage and citations are accelerated and
augmented by immediate OA compared to a one-year OA embargo will no
doubt vary by discipline, being greatest for fast-growing,
fast-turnaround fields; but it is hard to believe that there is any
field in which it is worth publishing one's results at all where it is
not worth making them accessible to all potential users immediately
upon publication.)

> From a more philosophical viewpoint I would like to raise the issue of
> weather each article reading is equally valuable from society's viewpoint. A
> very important type of reading is where the reader finds an interesting
> citation and tries to retrieve the cited article. For this type of reading
> 12 month delayed OA provides almost an equal service to full OA.

That's hardly true (or merely philosophical) if that reading happens
to occur within a year of the publication of the cited article (and
it's another reason for self-archiving articles immediately upon
acceptance for publication, rather than waiting out the publication
lag).

> And usually the chances are much higher that these readings influence the readers own
> research and that the article is read more carefully than the average
> current awareness reading where researchers quicly scan new articles in the
> journals they follow.

That sounds like a more philosophical point as it stands. It would be
more useful to have objective evidence on that too (for example,
either peer ratings or citation counts for articles that cite other
articles within different citation-latency periods)...

Stevan Harnad