[OAI-eprints] Recommendations for UK Open-Access Provision Policy

[Stevan Harnad]

    RECOMMENDATIONS FOR UK OPEN-ACCESS PROVISION POLICY
    To UK Government Science and Technology Committee

http://www.parliament.uk/parliamentary_committees/science_and_technology_committee/scitech111203a.cfm

>   "The Committee will be looking at access to journals within the
>   scientific community, with particular reference to price and
>   availability."

A more general way to put it would be "access to the articles published in
peer-reviewed journals". The articles (2,500,000 annually) are research
output. Researchers publish them in peer-reviewed journals (24,000
in all, across all scientific and scholarly disciplines, worldwide)
in order to make them accessible to all other researchers (worldwide)
to be read, applied, used, built-upon, cited: This is called "research
impact" and it is what is behind research productivity and progress (as
well as the career advancement and future research funding of the
researcher, the prestige and research funding of the researcher's
institution, and the benefits to the UK tax-payer for the money spent
funding the research).

>   "It will be asking what measures are being taken in government,
>   the publishing industry and academic institutions"

It is extremely important to separate the sectors over which the
UK government has some direct control -- government itself, and
academic institutions -- from the ones over which it can only have some
indirect influence: the publishing industry.

The UK government can do a great deal to maximise the access to and
the impact of UK research output through government research funding
policies and through HEFCE influence over academic institutional policy,
in particular, by extending existing publish-or-perish policy to mandate
open-access provision:

    UNIFIED OPEN-ACCESS PROVISION POLICY:
    (OAJ) Researchers publish their research in an open-access journal if
    a suitable one exists, otherwise (OAA) they publish it in a suitable
    open-access journal and also self-archive it in their own research
    institution's open-access research archive.

     Harnad, S., Carr, L., Brody, T. & Oppenheim, C. (2003)
     Mandated online RAE CVs Linked to University Eprint
     Archives: Improving the UK Research Assessment
     Exercise whilst making it cheaper and easier. Ariadne 35.
     http://www.ariadne.ac.uk/issue35/harnad/

http://www.ecs.soton.ac.uk/~harnad/Temp/self-archiving_files/Slide0022.gif
http://www.ecs.soton.ac.uk/~harnad/Temp/self-archiving_files/Slide0024.gif

But government can only influence publishers indirectly. The greatest indirect
influence will be the effect of the above open-access provision policy itself,
if it is mandated. This will encourage journals (first) to support author
self-archiving
http://www.lboro.ac.uk/departments/ls/disresearch/romeo/Romeo%20Publisher%20Policies.htm
and -- perhaps -- eventually also to become open-access publishers:
http://www.ecs.soton.ac.uk/~harnad/Temp/self-archiving_files/Slide0028.gif

Secondarily -- but note that the amount of open access to UK research this
will help provide is far less than the amount that will be provided by
the above open-access provision policy -- the government can also provide
(as part of research support) some support for the costs of publishing
in open access journals, to further encourage publishing in open access
journals, to help sustain the small number of open access journals that
exist today (600, vs. 23,400 toll access journals), and to encourage the
creation of new open access journals and the conversion of toll-access
journals to open access.

But note that the greatest impetus to this (possible eventual)
transition from toll-access publishing to open-access publishing
will come from mandating open-access *provision* itself (by the
joint OAJ/OAA route), for this will generate open access directly --
and *perhaps* eventually also the university journal subscription
cancellations from which the annual university windfall savings will be
the natural source out of which to pay the open-access journal publication
(peer-review) costs for each university's *own* research output:
http://www.ecs.soton.ac.uk/~harnad/Tp/resolution.htm#4.2

>   "to ensure that researchers, teachers and students have access to the
>   publications they need in order to carry out their work effectively."

The government can of course act for all of these good reasons. But
remember that most of peer-reviewed journal research is written by
researchers for researchers, to be used, applied and built upon in
further research, to further research progress. Otherwise it is hardly
read by anyone (including teachers and students).

So the government's open-access provision policy has to be very clear
both on why open access to this special literature is so important and
necessary (for the sake of research productivity and progress) and how it
can make this importance and necessity known to researchers, so that they
will want to support the mandating of open-access provision: Researchers
will support it for the sake of enhancing research impact. That they will
understand and approve fully. But they will not be much persuaded (and
perhaps even resistant) if they are told that open-access provision is
mandated in order (1) to encourage publishers to convert from toll-access
to open-access publishing, (2) to save money for libraries, (3) to provide
access to research for teachers, students and the general public, or even
(4) to provide access to research for the developing world.

(1) - (4) may all be valid reasons for the *government* to support
open-access provision, but for the *researcher* the only persuasive
reason is: to maximise the impact of his own peer-reviewed research output
(thereby maximising its contribution and benefits to science, as well
as the resulting rewards to the researcher and his institution).

>   "The inquiry will also examine the impact that the current trend towards
>   e-publishing may have on the integrity of journals and the scientific
>   process."

There is no "current trend toward e-publishing"! Virtually all of the
24,000 peer-reviewed journals are already hybrid: They publish both
an on-paper and an online version, both still accessible only through
institutional tolls. There are a few online-only journals, but these are
not necessarily open-access journals (of which there are about 600). So
do not confuse hybrid-online or online-only journals with open-access
journals.

All journals have benefitted from the economies and efficiencies of the
online medium for processing submissions, implementing peer review, and
producing and distributing both the paper and online edition. But
those economies and efficiencies themselves have not inclined most
journals to convert to open access.
(Only 600 out of 24,000 have done so to date.)

So the electronic medium itself has increased access for those
institutions that could afford the tolls, because licensed online
institutional toll-access provides more and better access than paper
subscriptions do. But the electronic medium certainly has not generated
open access -- far from it. It is still a fact for *every one* of the
2,500,000 peer-reviewed journal articles published annually that *most*
of its would-be users cannot access it, because their institutions cannot
afford the access-tolls. This means that an estimated 336% of potential
research impact has been lost, and continues to be lost, daily:

http://www.ecs.soton.ac.uk/~harnad/Temp/self-archiving_files/Slide0006.gif
http://www.ecs.soton.ac.uk/~harnad/Temp/self-archiving_files/Slide0007.gif
http://www.ecs.soton.ac.uk/~harnad/Temp/self-archiving_files/Slide0025.gif

    Lawrence, S. (2001a) Online or Invisible? Nature 411 (6837): 521.
    http://www.neci.nec.com/~lawrence/papers/online-nature01/

    Lawrence, S. (2001b) Free online availability substantially
    increases a paper's impact. Nature Web Debates.
    http://www.nature.com/nature/debates/e-access/Articles/lawrence.html

    Kurtz, M.J. et al. (2003) The NASA Astrophysics Data System:
    Sociology, Bibliometrics, and Impact. Journal of the American Society
    for Information Science and Technology
    http://cfa-www.harvard.edu/~kurtz/jasis-abstract.html

This has nothing to do with the "integrity of journals and the scientific
process." Journals are journals, whether paper or online, whether
toll-access or open-access. And the journal's component in the scientific
process -- the administration of peer-review (the peers review for free)
-- is unchanged, whether it is administered on paper or online, and
whether its administration costs are recovered on a toll-access publishing
cost-recovery model or an open-access publishing cost-recovery model. The
only thing that has been changed (and changed radically) by the advent of
the online medium is the possibility, at last, of providing open access
to this special literature that its authors have always given away for
free (even to the point of mailing hard-copy "reprints" at their own
expense to any would-be users who asked for them) in order to maximise
their research impact.

    Harnad, S. (1998/2000) The invisible hand of peer review. Nature
    [online] (5 Nov.  1998) and Exploit Interactive 5 (2000):
    http://helix.nature.com/webmatters/invisible/invisible.html
    http://www.exploit-lib.org/issue5/peer-review/

>   "What impact do publishers current policies on pricing and
>   provision of scientific journals, particularly big deal schemes,
>   have on libraries and the teaching and research communities
>   they serve?"

Separate the serials budget problem of university libraries from the
research impact problem of university researchers. They are related
and connected, but not in an obvious way, and they are certainly not
the same problem. 

Libraries must make do -- and provide access for their researchers to
whatever they can afford -- from year to year. For them, online licensing
has been a boon: more journal titles and articles accessible to more
of their institutional researchers, per pound paid in access-tolls.

But prices keep going up too. So there is also a shrinkage in the number
of journals libraries can afford. The "big deals" offer libraries the
bonus of getting both the paper and the online version of all journals
(from the same publisher) that they have subscribed to previously, plus
all journals (by the same publisher) to which they did not subscribe
previously -- for the price of only the journals they subscribed to
previously. 

This "big deal" too provides some increased access, but the prices still
keep going up. So the net outcome is the same: An overextended journals
acquisition budget (at the cost of an underfunded book acquisitions
budget) and affordable access to only a tiny fraction of the annual
2,500,000 articles in the 24,000 journals.

This means university libraries remain cash-strapped, and their users
remain access-deprived (not relative to what they used to have, in paper
days, but relative to all there is, the 2,500,000 annual articles in
the 24,000 peer-reviewed research journals): That's the serials budget
problem, and it is purely on the input/buy-in side. 

But there is also the research impact problem, which is on the output
side: University researchers are impact-deprived -- because of the access
problems of *other* universities: *those* universities cannot afford
access to *my* university's research output, so *I* lose research impact.

The two problems are connected, but in a subtle way. The key to
understanding the two problems is to understand the reciprocity
involved. Libraries tend to misunderstand and mis-state this as:
"Our university does the research, gives it away to publishers for free,
and then has to buy it back!"

This is completely incorrect. What the university is buying *in* (not
back) is the research output of *other* universities, not their *own*
research output. (They already *have* their own research output!) What is
being lost is research impact: the consequences of access-denial to *my*
give-away research because *other* universities cannot afford the tolls
to access the journal in which it appeared (hence cannot not read/use/cite
it).

The picture seems complicated, but the solution -- in the first instance,
to the lost research-impact problem, but eventually perhaps also to the
serials budget problem -- is to capitalise on the new online medium as
well as the peculiar reciprocity relation among the respective author
give-aways, by mandating that universities extend their publish-or-perish
policies to include open-access provision for those publications: It is
not enough to publish, and let the affordability of access-tolls determine
who can and cannot use your research. Publication must be supplemented
with open-access provision:

    UNIFIED OPEN-ACCESS PROVISION POLICY:
    (OAJ) Researchers publish their research in an open-access journal if
    a suitable one exists, otherwise (OAA) they publish it in a suitable
    open-access journal and also self-archive it in their own research
    institution's open-access research archive.

The result, in the short run, will be open access to all UK research output,
thereby maximising its research impact. 
http://www.ecs.soton.ac.uk/~harnad/Tp/resolution.htm#4.1

In the longer run this *might* also lead to a transition from toll-access
to open-access journal publishing, thereby solving the libraries'
serials budget problem.
http://www.ecs.soton.ac.uk/~harnad/Tp/resolution.htm#4.2

     "What action should Government, academic institutions and
     publishers be taking to promote a competitive market in
     scientific publications?"

Trying to increase between-journal competition in order to lower prices is only
a library serials-budget strategy. This has been going on for years now (led by
SPARC and SPARC-Europe, a consortium of university libraries trying to use their
collective consortial power to drive down journal prices). Its success
has been minimal, and its effect on researchers' access and impact has
been negligible.
http://www.arl.org/sparc/
http://www.sparceurope.org/

The reason this strategy does not work is because of inelastic demand for
peer-reviewed research. The 24,000 journals have a priority hierarchy
in this inelastic demand: All researchers need access to it all, but
no university can afford access to more than a fraction. So it is just
a matter of trying to buy in as much as each can, top-down.

The journals know (and feel, from the market's responses to price
increases) that the demand is inelastic: that the university libraries
have no choice. Moreover, because of the peculiar reward-structure of
this anomalous form of publishing -- unlike book authors, peer-reviewed
journal article authors *give away* their articles, seeking no royalties
or payment, but only research impact -- the only relevant competition
among journals is for *articles* (i.e., to get the best articles); there
is little competition between journals for subscriptions.

And competing for the highest-quality authors and articles depends,
paradoxically, on *rejecting* articles, in order to maintain the highest
standards of peer review. For it is the highest-quality articles that
generate the highest research impact (usage, citation).

So the top-down variable in the journal hierarchy is quality and
impact. This is the true determinant of what journals will and will
not be subscribed to by the libraries. It is for impact that journals
compete. But peer-review quality standards and rejection rates have
absolutely no connection with any competitivity one might generate
between journal subscription prices!

So the path of trying to spark competition between journals in order
to lower access tolls is one that has afforded and promises limited
success. SPARC has subsidized and offered consortial subscription support
to lower-price journals. It is now doing the same for open-access
journals. But the scope for any substantial change here is very
limited, and it concerns mainly the libraries' year-to-year serials
budget problems; it has little impact on the access problem, hence the
research impact problem. (600 open access journals out of 24,000 journals
represents a very small portion of actual and potential impact space).

The way to solve the research impact problem is:

    UNIFIED OPEN-ACCESS PROVISION POLICY:
    (OAJ) Researchers publish their research in an open-access journal if
    a suitable one exists, otherwise (OAA) they publish it in a suitable
    open-access journal and also self-archive it in their own research
    institution's open-access research archive.

>   "What are the consequences of increasing numbers of open-access
>   journals, for example for the operation of the Research Assessment
>   Exercise and other selection processes? Should the Government
>   support such a trend and, if so, how?"

It is not the (very slowly) increasing number of open-access journals
(OAJ) that is relevant to the Research Assessment Exercise (RAE), nor even
(the almost as slowly) increasing number of articles made open-access
by self-archiving (OAA). What is relevant to the RAE, and what the Government
should support, is increasing the amount of open-access provision --
via both OAJ and OAA -- by mandating it.

>   "How effectively are the Legal Deposit Libraries making available
>   non-print scientific publications to the research community,
>   and what steps should they be taking in this respect?"

This is irrelevant. Legal deposit libraries store copies of record,
for archival and preservation purposes. They are not open-access
providers. What should be mandated is that all universities make
their *own* published research articles openly accessible by
publishing them in an open-access journal and/or depositing
them in their own university open-access archives:
http://www.ecs.soton.ac.uk/~harnad/Temp/archpolnew.html

>   "What impact will trends in academic journal publishing have on
>   the risks of scientific fraud and malpractice?"

It is not clear why research access and impact problems are
being considered in the same breath with problems of fraud and
malpractice. There was some fraud and malpractice in paper journals. There
is some fraud and malpractice in online journals. The cost-recovery model
(whether toll-access or open-access) is also irrelevant. It is true that
it is easier to plagiarize or to otherwise misuse an online text than
a paper one, but it is also true that plagiarism and misuse are more
easily detectable online. So these balance out.

Apart from that, questions about scientific fraud and malpractice (and
questions about modifying the peer review system in any way) have nothing
to do with the question of open online access.

>   "In announcing the inquiry, the Chairman of the Committee, Ian
>   Gibson MP, said Journals are at the heart of the scientific
>   process.  Researchers, teachers and students must have easy
>   access to scientific publications at a fair price."

As noted, the access problem for this specialised literature is not
primarily a problem of teachers and students, but of researchers,
for the sake of research productivity, progress and impact. Nor is
it about *ease* of access: It is about having access *at all* (as opposed
to access denial). Nor is it merely or even primarily about having access
at "a fair price." This is an author give-away literature, written purely
for the sake of research impact. Access-denial at *any* price is already
needless impact-denial. Even if all 24,000 peer-reviewed journals were
sold *at cost* (and cost was minimised using all the economies and
efficiencies of the new electronic medium) it would *still* be true of
*all* 2,500,000 annual articles that they are inaccessible to most of
their would-be users, because their institutions still cannot afford
access to them all, and hence that all that potential research impact
is needlessly lost. 

The only remedy is to supplement toll-access (whatever its going price)
with open-access provision by the authors, institutions and funders that
provide this give-away research in the first place:

>   UNIFIED OPEN-ACCESS PROVISION POLICY:
>   (OAJ) Researchers publish their research in an open-access journal if
>   a suitable one exists, otherwise (OAA) they publish it in a suitable
>   open-access journal and also self-archive it in their own research
>   institution's open-access research archive.

>   "Scientific journals need to maintain their credibility and integrity
>   as they move into the age of e-publication. The Committee will have
>   some very tough questions for publishers, libraries and government
>   on these issues."

There is no credibility/integrity problem for the 2,500,000 articles
appearing annually in the world's 24,000 peer-reviewed journals. There is
an *access* problem for their would-be users -- those whose institutions
cannot afford the access-tolls -- and an *impact* problem for (all)
of their authors.

The tough questions should not be directed primarily at publishers and libraries
but at the research community itself: researchers, their institutions, and their
governmental research funders. And the question is: Why are the potential benefits
of this research not being maximised by maximising the access to it (through
open-access provision)? It is the research community that it in the position
to solve this problem -- especially of government mandates it.

Stevan Harnad