OA IRs Are Research Access Providers, Not Publishers or Library Collections

From: Stevan Harnad <harnad_at_ecs.soton.ac.uk>
Date: Tue, 4 Mar 2008 21:16:23 +0000

On Tue, 4 Mar 2008, John Smith, University of Kent, wrote:

> The problem is that the item referred to may have been published in a
> peer reviewed journal but the full text available from the repository
> may not be the final peer reviewed version.

John, I may be wrong, but I think that some lines may be crossing here,
between the immediate needs and concerns of researchers, trying to access
and use the current research literature, and the needs and concerns of
library cataloguing, concerned to collect and tag the canonical text.

Green OA is author self-archiving. If the author deposits a text in his
OA IR and tags it as "peer-reviewed" (and specifies the journal [date,
volume] that accepted it for publication), that is good enough for Green
OA, and for the immediate needs and concerns of researchers, trying to
access and use the current research literature (and particularly the major
portions of it to which they do not have subscribed institutional access).

Researchers have no pressing current need for a new certification
system, certifying that (1) the self-archiving author is telling the
truth that his self-archived draft was indeed peer-reviewed, and that
(2) the self-archived postprint is word-for-word identical with that
published article.

(Both this certification and this version authentification are technically
feasible, of course, but they are simply not worth the bother at this
critical time, when what is missing and urgently needed for 85% of
annual peer-reviewed research articles is not certification that they
have indeed been peer reviewed or that they are word-identical to the
published version; what is needed is the articles themselves!)

> Is this peer reviewed or
> not? Strictly it is not but it is the best copy we can provide of a peer
> reviewed item. Ideally we need to distinguish between the item status
> (peer reviewed) and the full text status (non-peer reviewed). As far as
> I am aware EPrints does not support this out of the box. Also it is not
> clear how one would represent this to a harvester program.

EPrints tags the self-archived item as peer-reviewed if the self-archiving
author tags it as peer-reviewed. The journal is the one that certifies
peer review. If a user's urgent need is for the certification, and not for
access to the item itself, then there is always the journal to turn to.

(Yes, some sloppy authors, sometimes, will not update their self-archived
unrefereed preprints, and instead simply re-label them as "peer-reviewed"
when the final draft is accepted. Scholarly practice will take care of
that sloppiness in due course. But what's incomparably more important
right now is to stay focussed on solving the real problem -- not yet
solved, and not even being conscientiously attended to -- which is getting
the authors to self-archive those preprints and postprints in the first
place! We are not working here to make a rich corpus spic-and-span for
a library collection catalogue: We are trying to enrich the impoverished
research corpus, so researchers can get on with their work!)

> This is related to a note I sent previously asking if the repository
> should be seen as a store (and delivery mechanism) or as a publication
> list. If it is a store (as indicated by the name 'repository') the
> associated metadata should describe the item contained which in many
> cases will be a non-refereed version of a refereed article.

The repository should be seen as a way for researchers to make their
findings accessible to and usable by all their would-be users, worldwide,
and not just those that can afford access to the publisher's proprietary
version, as subscribed to by their institution's library.

It is ever so important to remind ourselves that an institution's
OA IR is not a library collection, and its tags are not library card
catalogues.

IRs can certainly generate CVs and publication lists, but, as always,
the published items that are cited in those CVs and publication lists
are the *articles published in their respective journals*! A postprint
in the author's IR is not the publication itself. It is a supplementary
draft provided for access purposes, for those researchers whose
institutions cannot afford access to the publisher's proprietary version.

> Finally, there is a certain amount of academic snobbishness about
> peer-review, in many non-STM subjects the peer-reviewed article in not
> the main form of accepted publication and in others the first publication
> of new knowledge is in conference or working papers which may be later
> written up for journal publication. Any system that automatically says
> 'refereed=good', 'non-refereed=bad' is going to miss a lot of good
> quality material.

Who is saying 'refereed=good', 'non-refereed=bad'? The IR's tag is just
"refereed" and "unrefereed". For the rest, it's caveat emptor. If you
want to restrict your search and usage to research tagged 'refereed',
fine. If you want to broaden your search and usage, that's fine too.
Again entirely a matter for scholarly practise to decide.

The primary target literature for the OA movement, and OA IRs, is
peer-reviewed journal and conference articles (about 2.5 million per
year), because those are all, without exception, author give-aways,
written purely for the sake of researcher usage and impact, not for the
sake of royalty income and/or a prestigious hard-copy imprimatur. For the
disciplines that also rely on books, those books are also welcome in OA
IRs, but not many of them are likely to be deposited for the time being,
because they are decidedly *not* just author give-aways, written for
the sake of researcher usage and impact, not for royalty income and/or
a prestigious hard-copy imprimatur.

So whether or not to restrict search and usage to refereed content
is a matter for the user-scholars alone to decide, and whether or not
to self-archive their books alongside their articles is a matter for the
self-archiving author-scholars alone to decide. (It certainly cannot and
should not be mandated at this time!)

> It occurs to me that what we need is a post-publication quality
> indicator - otherwise known as a weighted citation-count :-) . Could
> we automatically include this in our repositories (taken from a central
> service?) or should we leave this to the search services?

I cannot follow this at all: The post-publication metrics concern the
*publication* itself and not just, or primarily, the version that happens
to be accessible in the author's IR! By all means add IR download counts
and their growth metrics to the growing spectrum of research usage and
impact
metrics, both before and after publication (EPrints IRs are already
doing this); and by all means couple IR-native metrics with global
harvested metrics, harvesting them back to the IR from citebase, google
scholar, citeseerx (and ISI and Scopus where permitted/licensed!).

But remember that the IR's primary function is to provide open access
to the institution's refereed research output, along with open access
metrics to provide feedback and incentives for self-archiving authors.
The search and usage of external users, however, will almost never be
at the local IR level; it will be at the global harvester/indexer level.

Stevan Harnad

> > -----Original Message-----
> > From: Repositories discussion list [mailto:JISC-
> > REPOSITORIES_at_JISCMAIL.AC.UK] On Behalf Of Stevan Harnad
> > Sent: 29 February 2008 18:07
> > To: JISC-REPOSITORIES_at_JISCMAIL.AC.UK
> > Subject: Re: Required and Desirable metadata in a repository
> >
> > Bill Hubbard is spot-on on the utility of am explicitly
> > searchable
> > field indicating whether or not an item has been peer reviewed.
> > The
> > EPrints software has such a tag.
> >
> > (It is only likely to be useful at a harvester level, as
> > individual
> > repositories (IR) are only likely to be searched for
> > institution-internal purposes. So this is a metadatum worth
> > displaying
> > for harvesters, and harvesters should set up in such a way as to
> > make
> > it possible to search on only the peer-reviewed items, if the user
> > wishes.)
>
> TEXT DELETED
>
> > Stevan Harnad
> >
> > On 08-02-29, at 12:10, Hubbard Bill wrote:
> >
> > > Dear Colleagues,
> > >
> > > Just picking up on Ian Stuart's question as to opinion on
> > "Required"
> > > and
> > > "Desired" metadata fields for eprints records.
> > >
> > > Could I ask colleagues how they view a "peer-reviewed" field?
> > >
> > > In terms of what users want, my own experience from talking to
> > > academics
> > > is that when faced with a mass of Open Access eprints the great
> > > majority
> > > have asked unprompted about how to search only within peer-
> > reviewed
> > > material.
> > >
> > > And for this facility we need to give services a peer-review
> > field,
> > > unless they start interpolating from other metadata features
> > like
> > > journal-title or somesuch.
> > >
> > > Copyright and peer-review (p-r) are the two topics that can be
> > > guaranteed to come up in academic discussions in relation to
> > > repositories: the first from their perspective as an author, the
> > second
> > > from their perspective as researcher/user.
> > >
> > > My strong suspicion is that most of those academics that haven't
> > asked
> > > about a p-r filter would want the feature before they used OA
> > material
> > > as a habitual source for research. Again, it may be that they
> > didn't
> > > ask
> > > because they assumed that it was all p-r, or, that it was all
> > non-p-r.
> > > (I have found repositories have a slighted reputation in some
> > quarters
> > > (often BioMedical) as being all referred to as "pre-print
> > servers").
> > >
> > > In terms of ingest, I think that the author is the best person
> > to know
> > > if their eprint has been p-r'd and that a peer-review tick-box
> > would be
> > > an acceptable additional task. Authors are generally pleased
> > that their
> > > article has passed p-r and would probably be happy about noting
> > that.
> > > As
> > > to how that information is recorded, that is another matter.
> > >
> > > Does this agree with other colleagues' experience? Is a p-r
> > field
> > > required to facilitate future use of the material?
> > >
> > > Regards,
> > >
> > > Bill
> > >
> > >
> > >
> > > > -----Original Message-----
> > > > From: Repositories discussion list
> > > > [mailto:JISC-REPOSITORIES_at_JISCMAIL.AC.UK] On Behalf Of Ian
> > Stuart
> > > > Sent: 21 February 2008 14:41
> > > > To: JISC-REPOSITORIES_at_JISCMAIL.AC.UK
> > > > Subject: Required and Desirable metadata in a repository
> > > >
> > > > [This is primarily a question for those involved in
> > repositories for
> > > > e-prints, but others may have interesting views]
> > > >
> > > > Within your own Repository, what [primarily metadata] fields
> > are
> > > > *Required* and what are *Desired*?
> > > >
> > > > If you were advising a fellow Institution about setting up a
> > > > repository,
> > > > what fields would you advise as *Required* and what are
> > *Recommended*?
> > > >
> > > > If you were to harvest[1] from a repository, what fields would
> > you
> > > > consider essential, and what would you consider helpful?
> > > >
> > > > Following on from that: if you were to harvest the Depot (or
> > even the
> > > > Intute Repository Search), how would you hope to identify[2]
> > deposits
> > > > that could be imported into your own Institutional Repository
> > > >
> > > > [1] This is where I come in: The depot will have a transfer
> > > > service, but
> > > > what to transfer?
> > > > [2] I've had loads of thoughts on this one, and they all seem
> > > > to spiral
> > > > and knit and knot and hide their threads, and not actually
> > > > conclude in
> > > > any meaningful way.... for me.
> > > >
> > > >
> > > > --
> > > >
> > > > Ian Stuart.
> > > > Developer for The Depot,
> > > > EDINA,
> > > > The University of Edinburgh.
> > > >
> > > > http://edina.ac.uk/
> > > >
> > >
> > > --
> > >
> > > Bill Hubbard
> > > SHERPA Manager
> > >
> > > SHERPA - www.sherpa.ac.uk
> > > RSP - www.rsp.ac.uk
> > > RoMEO - www.sherpa.ac.uk/romeo
> > > JULIET - www.sherpa.ac.uk/juliet
> > > OpenDOAR - www.opendoar.org
> > >
> > > SHERPA
> > > Greenfield Medical Library
> > > University of Nottingham
> > > Queens Medical Centre
> > > Nottingham
> > > NG7 2UH
> > > UK
> > >
> > > Tel +44(0) 115 846 7657
> > > Fax +44(0) 115 846 8244
> > >
> > > * * * * * * * *
> > >
> > >
> > > This message has been checked for viruses but the contents of an
> > > attachment
> > > may still contain software viruses, which could damage your
> > computer
> > > system:
> > > you are advised to perform your own checks. Email communications
> > with
> > > the
> > > University of Nottingham may be monitored as permitted by UK
> > > legislation.
>
Received on Tue Mar 04 2008 - 21:27:35 GMT

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:49:14 GMT