SV: [SIGMETRICS] Query about journal (not author) self-citation rates (fwd)

[Bonnevie Ellen]

In Introduction to Informetrics: Quantitative Methods in Library,
Documentation and Information Science, Elsevier Science Publishers B.V.,
1990 p. 260 the two authors, L. Egghe and R. Rousseau, explain the
calculation of journal self-cited and self-citing rates and make the
following interpretations. They claim,
1)      "A high self-cited rate (a journal's self-citations related to the
number of times it is cited by all journals) is an indication of a journal's
low visibility.
2)      "A high self-citing rate" (a journal's self-citations related to the
total number of references it gives) is rather an indicator of the isolation
of the field covered by the journal.

We see two measures pointing in different directions, or maybe it is two
sides of a coin. E. Garfield suggests in his 1974-comment (Journal Citation
Studies. XVII. Journal Self-Citation Rates-There's a Difference', Current
Contents, 52, December 25, 1974) that a combination and comparison of the
two will give an indication e.g. about the rank of a journal. Highly cited
journals tend to have smaller self-cited than self-citing ratios, vice versa
for less cited journals.
According to Garfield "high self-citedness may be related to
self-centeredness, specialization, or monopsony, a kind of reverse
monopoly."

Whether the measures are apt to group or to rank journals (cf. Loet
Leydesdorff's comment on 26. March) depends on the kind of analysis. Anyway
it should be given a try.

May I point to a recent article of mine? E. Bonnevie, A multifaceted
portrait of a library and information science journal: the case of the
Journal of Information Science. Journal of Information Science, 29(1), 2003,
p. 16. Here the pattern of self-citations is used among other measures to
characterize a journal. Both self-citing and self-cited rates are calculated
and compared to a few other journals.

Like H. Small I think that journal self-citations are mainly author-driven
and not editor-driven.

Ellen Bonnevie





-----Oprindelig meddelelse-----
Fra: Stevan Harnad [mailto:harnad at ECS.SOTON.AC.UK]
Sendt: 26. marts 2003 21:44
Til: SIGMETRICS at LISTSERV.UTK.EDU
Emne: Re: [SIGMETRICS] Query about journal (not author) self-citation
rates (fwd)


Thanks to Peter Suber for forwarding this from the FOS Forum:
http://www.topica.com/lists/fos-forum/read

>Date: Wed, 26 Mar 2003 14:19:05 -0500
>From: Garrett Eastman <eastman at rowland.harvard.edu>
>Organization: Rowland Institute at Harvard
>To: peters at earlham.edu
>Subject: Re: [FOS] Query about journal (not author) self-citation rates
>
>Eugene Garfield wrote about this in the seventies:
>Journal Citation Studies. XVII. Journal Self-Citation Rates -- ...
>http://www.garfield.library.upenn.edu/essays/v2p192y1974-76.pdf

This paper described a comparison made in the early '70's, of same-journal
citation counts, comparing the proportion of a journal's self-citations
to all incoming citations *to* that journal, i.e., self/in as well as their
proportion to all outgoing citations *from* that journal (self/out). There
is some interesting variation described, related to the journal's age
and hierarchical level, as well as its disciplinary range and whether
it is a review journal. The question of possible inflation to augment
the journal's citation impact factor is not mentioned.

Stevan Harnad