Re: Chronicle of Higher Education: Misunderstanding about the Evans & Reimer OA Impact Study

From: Stevan Harnad <amsciforum_at_GMAIL.COM>
Date: Tue, 24 Feb 2009 22:38:45 -0500

On Tue, Feb 24, 2009 at 4:04 PM, Klaus Graf
<> wrote:

      2009/2/24 Stevan Harnad <>:

> (Re: Phil Davis) No, E & R do not show that:

> "the vast majority of freely-accessible scientific
      articles are not
> published in OA journals, but are made freely available
      by non-profit
> scientific societies using a subscription model."

> E & R did not even look at the vast majority of
      freely-accessible articles,
> which are the ones self-archived by their authors. E &
      R looked only at
> journals that make their entire contents free after an
      access-embargo of up
> to a year or more.

Is there any empirical evidence that there are more
articles in the web than articles free after an embargo? There
is a
lot af free backfile access for TA journals. And even you
exclude that
you have to proof your assertion.

Clearly, date of publication as well as date of being made freely
accessible both have to be taken into account here, otherwise the
comparison becomes meaningless (since about 2.5 million articles are
being published per year, and even if every single one of them were
being immediately self-archived by their authors, there would always
be more cumulative post-embargo backfile articles). 

So there are two ways to reckon this: One way is within the year of
publication (say, 2009): Of the publication-year 2009 articles that
are freely accessible today (in 2009), are more of them (a)
publisher-provided post-embargo articles, (b) OA journal articles, or
(c) self-archived articles? (It's worth someone's trying to do the
count, but I think it's pretty sure which will be the winner,
considering how long journal embargoes are, how many journals there
are, how many OA journals there are, and how many self-archived
articles from non-OA journals there are, per year.)

A second way to reckon this is after an embargo elapses: say, year
2008 articles in year 2009. Again, the comparison is between (d)
non-OA-journal articles from 2008, in 2009, that are left in
closed-access, (a) non-OA-journal articles from 2008, in 2009, that
have been made freely accessible post-embargo by their publishers,
(b) OA journal articles from 2008 in 2009, and (c) non-OA-journal
articles from 2008, in 2009, that have been made freely accessible by
their authors. My bet is still on (c), but what is certain is that
Evans & Reimer took neither (b) nor (c) into account in their study.

The longer one reckons past the embargo, of course, and the larger
the proportion of journals that release their contents after an
embargo, the "sooner" embargoed articles will catch up and perhaps
eventually "win." But by that reckoning, research is the loser.
Because OA is primarily, and most urgently, about access to current,
ongoing research, not about backfile research. (That is one of the
reasons I had proposed in 2005 to make "immediate" and
"permanent" part of the definition of OA -- along with "freely
accessible online." By that definition, articles that were made
freely accessible online only after an embargo would not be "OA" at
all: OA would be tied to the publication date, and everything
released only later would be delayed access, not OA -- just as
articles that were made freely accessible for, say, only one
promotional week from the date of publication, and thereafter
accessible only for a fee, would not be "OA" either. But upon
reflection, redefining OA to exclude such potential abuses would have
overcomplicated things and OA was left, despite the absurdity, as
meaning "free online access," even if it was just a fleeting 1-day
state a century after publication!) 

The publication-year and one year post-publication-year counts and
comparisons would be worth doing, however, if someone has the time.
(I am too busy trying to get more mandates to deposit immediately
upon acceptance for publication, so that ongoing research can benefit
from at least 63% immediate OA plus 37% "Almost OA" despite publisher
embargoes, thanks to the "email eprint request" Button of which Mr.
Graf appears to be so unfond...)  
Stevan Harnad
Received on Wed Feb 25 2009 - 03:39:46 GMT

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:49:43 GMT