Re: Author's final draft and citing

From: Stevan Harnad <amsciforum_at_GMAIL.COM>
Date: Tue, 30 Sep 2008 10:29:31 -0400

    [ The following text is in the "UTF-8" character set. ]
    [ Your display is set for the "iso-8859-1" character set. ]
    [ Some characters may be displayed incorrectly. ]

This question has been raised many times and it has a simple, clear
and correct answer, in two parts:

(1) Do not conflate the question of what to CITE -- that is always the
canonical published work itself, if the work is published -- with the
question of what version of it you managed to ACCESS.

(2) If you cannot afford access to the publisher's proprietary
version, then you access the OA version deposited in the OA
Repository, but you always cite the published work (and,
preferentially, add the URL of the accessed version too).

That's it. The only two other minor details are:

(3) If the work is unpublished, or not yet published, you cite it as
unpublished, and, again, add the URL of the version that you accessed.

(4) The two reasons why it is vastly preferable that OA mandates
should specify that it is the author's peer-reviewed, accepted final
draft (the "postprint") that is deposited in the OA repository,
rather than the publisher's proprietary PDF is (4a) that far more
publishers endorse setting access to the author's deposited postprint
as OA immediately, rather than after and embargo, and (4b) PDF is the
least useful and functional format, for both human users and for robot
data-mining.

Some comments below:


 On 9/29/08, Delasalle, Jenny <J.Delasalle_at_warwick.ac.uk> wrote:

> I like to quote the Versions toolkit which mentions in a survey response
> that most academics prefer to cite the final, published version...

Of course, and so they should. But we are talking about what to do if
you cannot ACCESS the publisher's proprietary version, and the answer
is, access the author's OA postprint version -- but cite the canonical
published work, as always.

> whichever version they have read (p9:
> http://www.lse.ac.uk/library/versions/VERSIONS_Toolkit_v1_final.pdf)

Exactly.

> Whilst you're speaking to academics, you could survey them to ask what
> they would do...

Good question, but the right answer is, as always: "If the work is
published, I cite the published work." And the URL of the OA version
should be added to the citation, as the accessed version.

> But there is no evidence that I know of to indicate that anyone will
> cite any papers they have read in a repository.

There is abundant evidence that they cite them, as preprints, and once
published, as the published work. While only the unpublished preprint
is available, they cite that, as an unpublished preprint. As soon as
the paper is published, they cite the published version. While they
can only access the preprint or the postprint, users access that; if
they can access the publisher's proprietary version, they access that.

Henneken, E. A., Kurtz, M. J., Warner, S., Ginsparg, P., Eichhorn, G.,
Accomazzi, A., Grant, C. S., Thompson, D., Bohlen, E. and Murray, S.
S. (2006) ˙˙E-prints and Journal Articles in Astronomy: a Productive
Co-existence ˙˙ArXiv, Computer Science, cs.DL/0609126, 22 September
2006, in Learned Publishing, Vol. 20, No. 1, January 2007, 16-22
http://arxiv.org/abs/cs/0609126

> For those with no
> option (no subscription access), a bland looking draft of the article is
> better than not being able to read it at all, though.

Yes, but if it is published, the published version is still the one to cite.

> One more point to note here: what do we mean by a "citation"? Our
> academics are chiefly concerned with citations in journals that are
> indexed by Web of Science. But there are other kinds of citations: links
> from others' web pages and reading lists, and from papers in less
> prestigious journals or in disciplines not well covered by WoS and grey
> literature

This is mixing apples and oranges: A scholarly/scientific citation is
just that: The citation, by a scholarly work or another scholarly work
(usually text to text). This is true whether or not ISI happens to
index the work. Citations to and from non-ISI journals, as well as to
and from books, are all classical citations.

Web links, however, and reading lists are certainly scholarly impact
metrics, but they are not citations.

> that will help to raise the academic's profile in general
> and help in the sharing of scholarly knowledge, even if not raising
> their citations directly.

The perspicuous way to put this that citation counts are only one
among a multitude of potential scholarly impact metrics:

Harnad, S. (2007) Open Access Scientometrics and the UK Research
Assessment Exercise. In Proceedings of 11th Annual Meeting of the
International Society for Scientometrics and Informetrics 11(1), pp.
27-33, Madrid, Spain. Torres-Salinas, D. and Moed, H. F., Eds.
http://eprints.ecs.soton.ac.uk/13804/

Brody, T., Carr, L., Gingras, Y., Hajjem, C., Harnad, S. and Swan, A.
(2007) Incentivizing the Open Access Research Web:
Publication-Archiving, Data-Archiving and Scientometrics. CTWatch
Quarterly 3(3). http://eprints.ecs.soton.ac.uk/14418/

> Our academics are often worried that the content of an early draft does
> not reflect well on them (it varies across the disciplines), but the
> post-print/accepted version that we ask for is identical to the
> published one in its content.

Bravo: That is the optimal institutional self-archiving policy.

> The main issue we have is that our
> academics either never created an accepted version (aka post-print) or
> never kept one when they had it.

This is only a problem for older articles. For recent, current and
future articles, all authors have their final drafts, and those are
the primary targets of OA.

> Often, it seems, the draft is submitted
> and then revisions are discussed in e-mails and telephone conversations,
> so only the earliest draft ever existed, which is one that I can well
> understand them not being comfortable with sharing, even if they kept
> it. Indeed, some have written their submitted version in a publisher's
> template so there is publisher copyright content from the earliest
> draft.

The solution is simple: If the corrections were not incorporated in
the author's last draft, scrape them from the PDF proofs and convert
them back to text in the postprint.

This is all just obvious scholarly practice in the online era.

> In such circumstances, it is difficult to persuade an author that
> they need to create an accepted version for repository deposit,
> incorporating the changes later discussed: I don't know of a single
> author who has done that.

You need not. Leave it to scholarly practice. Unscholarly authors will
quickly realize that if they do not make sure their OA version is
correct, they will be incorrectly cited and quoted. Just make sure
they deposit their final drafts, and enjoin them to make sure they
contain the corrections, and leave the rest to them. The institution
is not the publisher of the work, just a provider of supplementary
access.

> So, you can easily argue against academic resistance on the points you
> raised, but you can't force them to deposit if they're too busy to
> consider it worth their while, if a suitable version never existed (I
> feel that publishers are disingenous here in allowing deposit of
> versions that don't exist!) or if they're happy to sign away their
> copyright and delete early versions.

Publishers have nothing to do with this question, which is simply on
of practical scholarly practice. Don't worry about it: individual
practice will catch up with evolving scholarly best practice in the OA
age.

> You can hope to show that those who
> do deposit are gaining reputation and citations over those who don't,
> whilst over time it will become accepted practice to keep/create
> suitable versions to deposit in the repository (even to negotiate
> copyright agreements). Or you can make the academics an offer they can't
> refuse: a mandate that is enforced... assuming that you have such powers
> :-)

The mandate should be adopted in any case. And the deposited postprint
should be the one used for all assessment purposes. Leave the rest to
scholars' good sense...


> From: Repositories discussion list
> [mailto:JISC-REPOSITORIES_at_jiscmail.ac.uk] On Behalf Of S Nieminen
> Sent: 22 September 2008 07:25
> To: JISC-REPOSITORIES_at_jiscmail.ac.uk
> Subject: Author's final draft and citing
>
> How have your academics reacted to the fact that it is often the
> un-paginated author's version that needs to be put in the repository
> instead of the pretty publisher's version?

Pagination is a red herring. In quoting excerpts, specify locus with
section heading and paragraph number.

> I'm having to speak to a
> number of people shortly and this will come up more and more. Some
> research staff are worried that the draft does not "look good" or that
> they won't get cited from papers that have not page numbers etc.

An OA postprint "looks" infinitely better than an inaccessible
publisher's PDF to would-users who cannot afford access. And they
certainly generate more citations (for the canonical, published work).

> Research seems to show a great increase in number of downloads for OA
> papers, however, are author draft versions getting cited more? How would
> this happen? Do people read the draft paper and THEN chase up the
> published version whether freely available or not?

No, they read and use the accessible draft and cite the canonical
published work.

Stevan Harnad
Received on Tue Sep 30 2008 - 15:32:14 BST

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:49:28 GMT