German OA survey: Deutsche Forschungsgemeinschaft (DFG)

From: Stevan Harnad <>
Date: Wed, 11 Jan 2006 16:13:35 +0000

     "The Deutsche Forschungsgemeinschaft (German Research Foundation)
     is the central, self-governing research organisation that promotes
     research at universities and other publicly financed research
     institutions in Germany.

    "The DFG serves all branches of science and the humanities by funding
    research projects and facilitating cooperation among researchers.

    "DFG Publishes Study on Open Access

    "(9 January 2006) The dissemination of research findings on special,
    free-to-user online platforms (open access) is approved by a majority
    of research scientists in all disciplines. However, only around ten
    percent have used this method to publish their own articles. These
    are the results of a DFG study, published in German in July 2005
    and now available in English. The study can be downloaded from the
    DFG website."

The survey is the full analysis and report of results that were heralded
at the Berlin 3 meeting in Southampton:

The DFG survey corroborates the findings of two JISC/Key Perspectives
Surveys announced and discussed in this Forum:

There is in the DFG survey and the report of the results a bit of
confusion among "OA publication" ("gold"), "OA platform" (not clear
what that means), and "Secondary publication" (a misnomer for
the OA self-archiving of published postprints or their preprints)
("green"). But the concrete findings are clear, and confirm KP/JISC.

The DFG Survey finds that in 2004 there was more doubt in researchers'
minds about the quality of OA journals. Today that doubt has clearly
shrunk, as OA journals demonstrate both their quality and their capacity
to achieve high impact.

The DFG survey (like many such surveys) focuses disproportionately
on OA publishing (gold), and not enough on OA self-archiving (green),
nor on the all-important actual findings and implications of the survey
regarding green, which were: Three times as many researchers provide
OA to their articles via green as via gold (but that still only amounts
to <30% vs. <10%).

Yet over 70% of researchers are favorable to self-archiving. So the
natural follow-up question to have asked would have been: Why aren't you
*doing* it then?

The KP/JISC surveys did ask this, and got the answer: Most researchers
are too busy and set in their ways to self-archive systematically until
and unless it is required of them by their employers and/or funders. (But
if/when required, they will happily comply)! (Yet even this is now old
news, because these subjective judgments have since been objectively
tested in action -- by four institutions that actually implemented a
policy of requiring self-archiving -- Southampton, QUT, Minho & CERN --
and the result was exactly what the KP/JISC respondents had predicted:
self-archiving rates approaching 70%+.)

The DFG data on discipline differences are correct, but not
coherently interpreted: Yes, the humanities are more book-based than
journal-based. We already knew that. What follows form this fact? That
we should get the journal part self-archived (in humanities and the other
disciplines too) and then we can discuss other forms of content. Otherwise
the difference is simply a red herring. The primary OA target today is
journal articles. No discipline fails to benefit from enhanced
usage and impact for its journal articles. Books are moot for now.

Least informative are the opinions of the DFG respondents (70-90% of whom
have not yet self-archived, with many of them hearing about it for the
first time from the DFG survey itself) as to what the "optimal" form of
self-archiving would be: institutional or central. Researchers' opinions
can be duly noted, but then we have to remember what OA self-archiving
is about and for, which is about increasing the current low levels
of self-archiving (c. 15%) to 100% as soon as possible.

For that we don't need researchers' introspections about whether central
or institutional archiving would be nicer (those are mostly based on
the user- rather than the depositor-perspective anyway, and even that,
based on naive, unexamined assumptions about what "central" really means,
in the OAI interoperable age). What we need instead is data on which
form of archive is more readily created and, crucially, filled!

Here again, it is the KP/JISC analysis, in another study, of the pro's
and con's of central vs. institutional self-archiving that weighs the
factors clearly and explicitly.

And since then we also have the success of the 4 institutional archives
(Zurich still too early to say), which, it cannot be stressed enough,
can only be measured in relation to their full annual target content:

QUT and Minho and CERN and Soton/ECS success has to be measured against
QUT and Minho and CERN and Soton/ECS total annual research article output,
and hence Arxiv and PubMedCentral success has to be measured exactly the
same way: Against the annual research article output of its *total* annual
target content (all institutions' physics and biomedical research output,
respectively). A central archive has the advantage of a bigger numerator
than any individual institution, but it also has the disadvantage of a
far, far bigger denominator. (Arthur Sale's excellent analysis is the
most pertinent one here.)

So it is not about how authors and users imagine *search* space being
structured, in their fantasies about OA content, but how *deposit*
space needs to be structured in order to get *filled*, as quickly and
fully as possible, with its target OA content.

Hence the DFG survey confirms that in 2004 in Germany knowledge of
OA was already growing, both about OA green and about OA gold, and
so too was the practice -- about 3:1 -- and especially the sense that
OA is a "good thing" (70%+). Yes, humanities scholars publish more in
books than in journals (about 4:1) and yes, we would all like to search
and find content aggregated by subject. But it does not follow that
OA is less beneficial for journal content in the humanities than for
journal content in any other field; and the fact that we want to search
"centrally" by subject does not mean we should *deposit* centrally in
the OAI-interoperable/harvestable age.

What is needed is institutional self-archiving (plus central harvesting)
and institutional and funder mandates to ensure that the institutional
archives are promptly filled, as QUT and Minho and CERN and Soton/ECS's
are being.

Stevan Harnad
Received on Fri Jan 13 2006 - 03:24:50 GMT

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:48:11 GMT