Revision of Impact Factors

eackerma eackerma at UTKUX.UTCC.UTK.EDU
Mon Sep 18 10:51:38 EDT 2000

Excellent response! I amend my previous comments to apply only to the
discussion of author self-citations in the context of efforts to measure or
evalutate research performance or productivity.

As an aside, I wonder if we as bibliometricians, in our rush to change the
two-year citation window used by ISI to calculate the Impact Factor, have
given due consideration to the potential lost of comparability? For all its
faults, real and imagined, at least the ISI Impact Factor as currently
calculated  provides some sort of standard that allows comparability between
the results of various studies. The more special cases, ad hoc adjustments,
and other  individualized tinkering that is done to the Impact Factor's
calculation, the less comparable the end results are. This is particularly a
problem when one is using citation anlysis as part of a larger study examining
research performance and productivity, which to have meaningful results,
involves some kind of comparative measures. All of which, of course, leads to
the broader issue of a lack of some sort of standardized bibliometric measures
in general...

Eric Ackermann
School of Information Sciences
University of Tennessee-Knoxville
eackerma at

>> Date:    Fri, 15 Sep 2000 13:39:25 -0400
>> From:    eackerma <eackerma at UTKUX.UTCC.UTK.EDU>
>> Subject: Revision of Impact Factors
>> While I find the current discussion of ISI's Impact Factor quite
>> interesting,
>> I'm afraid that I will have to take issue with the (apparently)
>> widely
>> accepted notion that all self-citations should be omitted from any
>> impact considerations........................
>> Therefore, until there are published studies in the literature that
>> empirically demonstrate the necessity for doing so, removing all
>> self-citations from the record before conducting a bibliometric
>> evaluation of
>> research performance seems to be an unnecessary activity. It only
>> seems to add
>> more work to the task of citation analysis for no good reason,
>> while unfairly
>> penalizing researchers in newer, highly specialized, or currently
>> unpopular
>> fields.
>> Eric Ackermann
>> School of Information Sciences
>> University of Tennessee-Knoxville
>> eackerma at
>> ------------------------------
>In response to the beginning and the end of Eric Ackerman's message,
>I would comment as follows.
>The selfcitations are of two kinds. (i) When authors cite their own
>published work, irrespective of the source journals, and (ii) when
>there are citations to the same journal in articles published in it.
>Let me call (i) "authors selfcitations", and (ii) "journal
>(i) In case of authors selfcitations one must certainly allow for
>autors' necessity to "keep track" of her/his earlier publications.
>But certainly there must be a limit to it. (Incidentally - is there
>any analysis about the context type of authors selfcitations?) As far
>as something in one's last paper has to be backed by explanations
>already given in earlir publications - the authors selfcitations are
>clearly indispensible.
>However, once decided to use citation counting to  e v a l u a t e
>individual   s c i e n t i f i c  input to the world knowledge, the
>aim is to find out to what extent that individual's work has been
>referred to by his/her PEERS in THEIR publications. (Some like to
>name this "impact", I don't.) Obviously, the authors selfcitations
>should be excluded for evaluative purposes.
>(ii) Journal selfcitations are of interest in journals  e v a l u a t
>i o n   (studies), a topic of particular importance to which I shall
>come in another comment.
>Suffice here to say that while journal selfcitations should NOT be
>excluded, such data, recorded specifically as kind of a
>"claustrophobic" index, ought to be analysed in comparing journals,
>bearing in mind, of course, that there are new or exotic fields of
>research with a very limited number of source journals.
>Yours in discourse,
>Do You Yahoo!?
>Yahoo! Mail - Free email you can access from anywhere!
>End of SIGMETRICS Digest - 15 Sep 2000 to 17 Sep 2000 (#2000-132)

More information about the SIGMETRICS mailing list