Open Access News

News from the open access movement

Tuesday, January 12, 2010

OA across the federal government, hold the mandate

The Scholarly Publishing Roundtable --a US group consisting mostly of librarians, publishers, and provosts-- today released its recommendations on OA for publicly-funded research.  The group's "core recommendation" calls for OA, and calls for it across the federal government, but stops short of calling for an OA mandate:

Each federal research funding agency should expeditiously but carefully develop and implement an explicit public access policy that brings about free public access to the results of the research that it funds as soon as possible after those results have been published in a peer?reviewed journal.

Here are the group's eight specific recommendations:

  1. Agencies should work in full and open consultation with all stakeholders, as well as with OSTP, to develop their public access policies.
  2. Agencies should establish specific embargo periods between publication and public access. An embargo period of between zero (for open access journals) and twelve months currently reflects such a balance for many science disciplines. For other fields a longer embargo period may be necessary.
  3. Policies should be guided by the need to foster interoperability....
  4. Every effort should be made to have the version of record (VoR) as the version to which free access is provided. If the VoR is not included in a public access database, the article version or reference that is
    included should contain links back to the VoR on the publisher’s site.
  5. Government agencies should extend the reach of their public access policies through voluntary collaborations with nongovernmental stakeholders...[such as] publishers, universities, and other entities husbanding the results of research, within and beyond the U.S.
  6. Policies should foster innovation in the research and educational use of scholarly publications.
  7. Government public access policies should address the need to resolve the challenges of long?term digital preservation.
  8. OSTP should establish a public access advisory committee....

One of the group's background principles is that "the results of research need to be published and maintained in ways that maximize the possibilities for creative reuse and interoperation among sites that host them."  You don't have to squint too hard to see that as an endorsement of libre OA.

Youngsuk "YS" Chi, who became CEO of Elsevier's Science & Technology (S&T) Division last month, toward the end of the group's deliberations, did not sign the final report:

Primarily, I have a fundamental concern that the report supports an overly expansive role of
government and advocates approaches to the business of scholarly publishing that I believe are
overly prescriptive.

Mark Patterson, the Director of Publishing at PLoS, did not sign the report either:

The result is a set of recommendations that will significantly improve the currently limited access to federally funded research, but stops far short of recognizing and endorsing the opportunities to unleash the full potential of online communication to transform access to and use of scholarly literature.


  • There's a lot to like here:  the endorsement of OA at every federal funding agency, the endorsement of OA for the published editions of peer-reviewed articles, and the endorsement of libre OA and not merely gratis OA. 
  • The group neither calls for an OA mandate nor offers an argument against OA mandates.  It should have done one or the other.  By doing neither, it appears to ignore both the arguments in support of mandates and the evidence of their effectiveness.  If it weren't for this odd omission, the report would come very close to endorsing FRPAA, or even a stronger, libre version of FRPAA.  But FRPAA would require agencies to provide OA, even if, like this report, it would leave them free to go their own ways on most other policy details.  Some members of the group didn't like the way the NIH enhances the articles on deposit in PubMed Central; but that doesn't explain the group's skittishness about OA mandates.  (Clearly an agency could require OA to the author manuscripts or published editions without further enhancement.)  Even though PMC is OAI-compliant, some thought PubMed Central could do more to be interoperable with "external databases"; but that doesn't explain the skittishness about OA mandates either.  (Clearly an agency could mandate OA in a repository as interoperable as the technology allows.)  Beyond this, the group does not criticize the NIH policy, and it does not criticize FRPAA at all, leaving us in the dark about what, if anything, might be wrong with these mandatory approaches to OA for publicly-funded research.
  • I can't explain why the group didn't offer an argument against mandates, once it decided not to endorse them.  But there are some clues in the text about why it decided not to endorse them.  One is that it wants OA to the Version of Record (the published edition), and it knows, and often remarks, that a funding agency cannot mandate OA to the published edition.  Another clue is that it seems to prefer that the OA texts be hosted by publishers at their own sites, even though it would allow them to be hosted by government or institutional repositories.  But even together, these don't address, let alone outweigh, the arguments for mandates, and we all would have been better served by an argument whose premises and reasoning were open to inspection.
  • The problem with publisher-hosted OA is that it's uncertain.  Just as a funding agency can't compel publishers to provide OA to the published edition, it can't compel them to provide OA from their own sites.  In the absence of a binding policy, publisher willingness to host OA copies would be contingent, and highly variable.  (As a sign of the passive-aggressive forgetfulness we might face:  Just last month, Heather Morrison pointed out that Wiley-Blackwell's Health and Social Care in the Community has agreed to deposit articles by NIH-funded authors in PMC, but has a compliance rate of exactly 0%.)  [See the second update below.]
  • I'm one who agrees that the published edition is generally more useful than the final version of the author's peer-reviewed manuscript, although I'd add:  unless the published edition is only available in PDF.  But even if all the editions we're talking about are in HTML or XML, assured OA to the final version of the author's peer-reviewed manuscript is far more useful to research than untrustworthy (flaky, selective, temporary, late) OA to the published edition.  If we can have assured OA to the published edition, and in a use-friendly format, wonderful; I want it.  But if we can't, we should put assured OA ahead of OA to the published edition. 
  • One more clue as to why the group may have decided not to endorse OA mandates:  The only working scientists on the panel were either in library science or had become publishers.  I respect all the participants, but the panel omitted a significant, probably the most significant, stakeholder group.

Update.  In my message posting news of the report to the SPARC Open Access Forum, I mistakenly said that Elsevier and PLoS did not sign the final report.  I should have said that YS Chi and Mark Patterson did not sign the final report.  The members of the panel agreed to participate as individuals, not as representatives of their employers.  I regret the error.

Update (1/19/10).  I was wrong to criticize Wiley-Blackwell's Health and Social Care in the Community (HSCC) for not following its own policy to deposit articles by NIH-funded authors in PMC.  In the period since the NIH policy became mandatory, HSCC has had two submissions based on NIH funding.  In the first case it deposited the manuscript in PMC within six days of receipt.  The second paper was received very recently and is still in process.  (Thanks to Cliff Morgan for the correction.)  My apologies to HSCC and Wiley-Blackwell. 

Heather Morrison has also posted a correction.