The usage and respectability of webmetrics depend on our continuing
to analyse and demonstrate their reliability and validity. It is a
foregone conclusion that they will prevail. What is missing, along with
data and advocacy, are the following two critical complonents, in order
of importance:
(1) Content: Until/unless the OA movement gets the remaining
85% of the full-text content up there, we are just talking about
dilettantism
(2) Benchmarks and baselines: Absolute numbers mean absolutely
nothing; we need to develop comparative norms so we can give numbers
*relative* meaning too ("compared to what?"). (In other words,
usage stats need to be federated and compared against one another,
and against ongoing norms, not just floated in vacuo; for that too,
more content, and not just more stats, are urgently needed.)
Google Scholar (and Google Books too) are being short-sighted in focussing
mainly on the *vendors* of scholarly content (publishers). Universities
and researcher institutions, and in particular, their research authors,
are the primary *producers* of that content, and it is the growing
network consisting of their own IRs that is in a position to be its
primary online access-providers.
http://archives.eprints.org/
I am still waiting to hear from Anurag Acharya... If Google
collaborated with us, we could accelerate things substantially.
"Google's Scholarly Search Service and Institutional OA Self-Archiving"
(Nov 2004)
http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/4176.html
(Nor would it hurt if Arxiv at long last agreed to let us use their
US download stats, so we didn't have to rely on only the stats from
the UK Arxiv mirror we have been hosting at Southampton for nearly
a decade:
http://uk.arxiv.org/ )
"Early Download Impact Predicts Later Citation Impact" (Sep 2004)
http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/3951.html
http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/4387.html
Chrs, Stevan
On Tue, 17 Jan 2006, Leslie Carr wrote:
>
> On 17 Jan 2006, at 12:23, Stevan Harnad wrote:
>
> > PS I actually think there are a lot more things that can be counted
> > and done with usage stats than the standard ones,
>
> When Stevan started the JISC/NSF OpCit project in 1999 (in
> collaboration with Cornel and arXiv), we started to make use of
> download statistics in the Citebase service and demonstrated the
> correlation between downloads and citations (hardly surprising in
> retrospect).
http://citebase.eprints.org/analysis/correlation.php
http://arxiv.org/pdf/cs.IR/0503020
> At the time, download figures as usage data were very much mistrusted
> as shown in the following equation:
>
> web logs = web [global trivia] + logs [unreliable]
>
> I found it very uncomfortable when trying to present these results to
> an audience that only believed in citations and impact factors and
> considered Web stats to be vanity on top of unreliability.
>
> But now, usage stats are becoming more and more commonly used in
> digital library environments and repositories. The COUNTER
> organisation uses them to provide libraries with commercially-
> sensitive information. Herb vdS has recently presented a paper that
> proposes an OAI-based system that incorporates usage and citation
> statistics. I don't believe that Citeseer or Google Scholar make any
> use of usage stats (although the could capture the usage from their
> respective portals).
>
> So here's my question: are usage stats academically respectable yet?
> Individual researchers seem to respond well to the news that they
> have had a significant number of downloads, but I'm not sure that
> they treat the information as crucial. One could take the
> perspective that downloads (as indicators of reading) are a more
> natural form of "research takeup" evidence than citations. Perhaps
> citations are used because they have been the only visible audit
> trail of research impact.
>
> ---
> Les Carr
Pertinent Prior AmSci Topic Threads:
"Scientometric OAI Search Engines" (Aug 2002)
http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/2238.html
"How to compare research impact of toll- vs.
open-access research" (Jun 2003)
http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/2859.html
"On the Strong Causal Connection Between Access and Impact" (Apr 2004)
http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/3697.html
"Early Download Impact Predicts Later Citation Impact" (Sep 2004)
http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/3951.html
http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/4387.html
"OA advantage = EA + AA + QB + OA + UA" (Sep 2004)
http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/3978.html
"Self-Archiving Incentives: Download Impact Counts" (Oct 2004)
http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/4085.html
Received on Tue Jan 17 2006 - 16:09:10 GMT