Re: OA and impressive impact factors - non propter hoc

[Matthew Cockerill <matt@biomedcentral.com>]

Tony, It's great that Elsevier is taking an interest in BioMed Central
journals and their impact factors. I do take issue with your conclusions,
however.

Firstly, the statistics you give are based on the set of journals that
have ISI impact factors (in fact, they cover only journals which had 2003
Impact Factors). Unfortunately, ISI's tracking of new journals is patchy
and inconsistent. Many of BioMed Central's best journals are not yet
tracked by ISI. We are able to calculate 'unofficial impact factors' for
journals not tracked by ISI, and these confirm that many of these
untracked journals are cited as frequently (or more) as our journals that
are ISI tracked.  Since only a rather arbitrary subset of our journals is
tracked by ISI, attempts to make quantitative comparisons in terms of
number of titles are likely to be misleading.

Secondly, comparing the percentage of Impact Factors going up or down does
not seem a particularly meaningful metric . What is important, surely, is
the actual value of the Impact Factor (relative to others in the field).
In that regard, BioMed Central titles have done extremely well, and
several are close to the top of their disciplines. For example, BMC
Bioinformatics only received its first Impact Factor this year (5.42), and
so was ignored by your 'Impact Factor change' metric.  But the fact
remains that this is an excellent Impact Factor which puts it right at the
top of its field. It is, in fact, only a whisker behind the
best-established journal in the field (Bioinformatics from OUP), which has
an Impact Factor of 5.74

Thirdly, you raise the point that review articles can boost a journal's
Impact Factor, and that many journals publish review articles specifically
with the intention of improving their Impact Factor. This is certainly
true, but of BioMed Central's 130+ journals, all but six are online
research journals, and publish virtually no review articles whatsoever. In
that light, the achievement of journals such as BMC Bioinformatics is even
more impressive, as they have achieved their Impact Factors purely on the
basis of their research content, while many of the journals they are
competing against publish reviews.  Perhaps ISI should be encouraged to
calculate a separate 'Research Impact Factor,' excluding reviews. This
would eliminate at least one source of bias from Impact Factor
comparisons. It would be very interesting to see how this would affect the
relative positions in ISI rankings of titles from BioMed Central vs other
publishers .

Best regards,

Matt Cockerill
Director of Operations
BioMed Central Limited

On 8 Jul 2005, at 15:39, Mcsean, Tony (ELS) wrote:

> To assess the full significance of improved impact factors (IFs) for
> many BMC titles, and in particular to assess the open access factor, we
> need to put the BMC results into context with those of the major
> subscription-based publishers. Elsevier's research staff has analysed
> the figures and come up with the following results:
>
> Of the BMC journals that have IFs for 2004 and 2003, 60% have increased
> and 40% declined (in actual numbers of journals - of those that had IFs
> in 2003, 11 increased and six declined.
>
> The equivalent results for Elsevier, Wiley, OUP and AIP show:
>
> 	Elsevier - 64% increased IF, 36% decreased IF
> 	Wiley    - 58% increased IF, 42% decreased IF
> 	OUP     - 64% increased IF, 36% decreased IF
> 	AIP       - 65% increased IF, 35% decreased IF
>
> In addition, not all of the content of BMC journals is based on the
> author-pays open access model but is available on a traditional
> subscription basis. The Review, Viewpoint and Commentary articles of
> some journals fall into these categories. Typically these article types
> accumulate more citations then standard research papers (see:  
> http://www.elsevier.com/framework_editors/pdfs/Perspectives1.pdf).
>
> For example, between 50% and 65% of the content of Arthritis Research &
> Therapy, Critical Care, and Breast Cancer Research is made up of review,
> viewpoint or commentary articles. These subscription-based articles,
> particularly review papers, account for a larger proportion of the IF of
> these journals then do the author-pays open access research papers.
>
> This comparison with four major STM publishers demonstrates that BMC's
> overall IF results are unremarkable, and that they certainly do not
> provide evidence to support the common assertion that the open access
> publishing model increases impact factor scores.
>
> Tony McSeán
> Director of Library Relations
> Elsevier
> +44 7795 960516
> +44 1865 843630