Re: Open access and impact factor

[Stevan Harnad <harnad@ecs.soton.ac.uk>]

Rick Anderson asked:

>   [I]sn't open access simply making the impact-factor data less
>   meaningful? ...given two articles of equal merit and potential
>   influence, one of which is freely available to the public and the
>   other of which is only available to those who pay, wouldn't we expect
>   that the impact of the former would be higher than that of the latter?
>   And if so, how is the difference between those two impact factors
>   meaningful or useful?

The reason researchers publish their research ("publish or perish") rather
than just putting it into a desk-drawer and moving on to do the next piece
of research is so that it will be used, so that it makes a contribution to
knowledge, so it makes a difference. This is research impact, of which
citation counts are one correlate, hence predictor.

Findings that cannot be accessed by all their would-be users are findings
that cannot have their full potential research impact.

The purpose of open access is to maximise the visibility, usage and hence
the impact of research. To maximise the visibility, usage and impact of
one's own research by providing open access to it is not to make the
impact factor less meaningful or more meaningful. It is to maximize the
impact of one's own research, just as publishing it at all was done in
order to make it visible and usable, so the work can have its impact
rather than remaining still-born, done in vain.

Today, most journal articles are toll-access (TA), hence impact-limited,
rather open-access (OA).  How do we get there (universal OA) from here
(restricted access)? By providing OA to one's own articles. Yes, of
course, in the unlikely event that two articles are of exactly equal
merit, and one is TA and the other OA, the OA article will have more
users. That's the point! And the remedy for this momentary imbalance is
that the OA should be provided for the other article too, and then access
to both will be be maximized, and merit alone will again rule.

The race is to the swift. Once all articles are OA, their impacts will all
be rescaled upward, because all articles will have maximized impact. Until
then, the OA article will have an extra edge over the TA article. Let that
serve as another incentive to provide OA as swiftly as possible!

A word about "journal impact factors," though: As David Goodman and
Jean-Claude Guedon pointed out, these are merely average citation counts
for the articles in a journal. That's a blunt instrument. The sharper
instrument is the article's own citation count (and the author's, and the
download count, which is an early-days measure, and highly correlated
with, hence predictive of, later citation counts).

OA will not only give us a far richer array of impact indicators, but it
will put the focus on the individual article's merit, rather than only on
the average merit of the articles in the journal it was published in. (Do
not, however, underestimate the predictive value of the journal name
itself, for the journal is the peer-review service-provider, and its track
record for the average quality of its articles is also an indicator of its
acceptance standards. There are 24,000 journals, differing not only in
field, but in peer-review standards and selectivity; there is usually a
journal hierarchy of merit within each field. This filter is not to be
sneezed at either.)

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 (April 2003).
http://www.ariadne.ac.uk/issue35/harnad/

Harnad, Stevan (2003) Measuring and Maximising UK Research Impact. Times Higher
Education Supplement. Friday, June 6 2003.
http://www.ecs.soton.ac.uk/~harnad/Temp/thes.html

Harnad, Stevan (2003) Maximising Research Impact Through Self-Archiving.
http://www.ecs.soton.ac.uk/~harnad/Temp/che.htm

Stevan Harnad