Tibor Braun, Wolfgang Gl änzel and András Schubert "A Hirsch-type index for journals" The Scientist Vol:19 #22 p.8. November 21, 2005

[Eugene Garfield]

Tibor Braun : braun at mail.iif.hu

TITLE : A Hirsch-type index for journals
AUTHOR: Tibor Braun, Wolfgang Glänzel and András Schubert
SOURCE: The Scientist Vol:19 #22 p.8. November 21, 2005

Table 1: Journals with the Higest H-Index for  their 2001 Papers

Rank by   Journal Title            Journal h-index       Rank by 2001
h-index                                                  Impact Factor
_______   ______________           ________________      ______________

1       Nature                            157                  10
2       Science                           155                  13
3-4     New England Journal of Medicine   113                   5
3-4     Proceedings of the National       113                  55
        Academy of Sciences (USA)
5       Cell                              109                   3
6       Journal of Biological Chemistry   100                  95
7       Physical Review Letters            96                 118
8       Lancet                             89                  60
9       Circulation                        86                  54
10      Nature Genetics                    85                   4


SOURCE: Web of Science, accessed September 16, 2005

Re: the h index.1 We suggest that a h-type index – equal to h if you have
published h papers, each of which has at least h citations – would be a
useful supplement to journal impact factors. First, it is robust and
therefore insensitive to an accidental excess of uncited papers and also to
one or several outstandingly highly cited papers. Second, it combines the
effect of "quantity" (number of publications) and "quality" (citation rate)
in a rather specific, balanced way that should reduce the
apparent "overrating" of some small review journals.

The journal h-index would not be calculated for a "lifetime contribution,"
as suggested by Hirsch for individual scientists, but for a definite
period – in the simplest case for a single year. Fortunately, the Web of
Science database offers a very simple way to determine the annual h-index
of a journal without the need for any off-line data processing. Retrieving
all source items of a given journal from a given year and sorting them by
the number of times cited, it is easy to find the highest rank number which
is still lower than the corresponding times cited value. This is exactly
the h-index of the journal for the given year.

We chose 2001 as a starting year, and looked for citations until the time
of accessing the database on September 16, 2005. We used the Journal
Citation Reports 2001 for comparative impact factor data. The list of the
ten journals with the highest h-index for their 2001 papers is given in
Table 1. Conspicuously, the first and second ranked journals of the 2001
impact factor list – the Annual Review of Immunology and the Annual Review
of Biochemistry – are missing from the table. Since they published 24 and
23 papers, respectively, in 2001, they had no chance to compete with the
chart toppers, because the h-index cannot be larger than the number of
papers it is based on.

This in no way meant to belittle the significance of these journals, but
does stress the different dimensions emphasized by the two indicators. Not
surprisingly, the majority of the journals in the table below are from the
biomedical field, a fact that underlines the necessity of discipline-
specific evaluation of this indicator, as well. Nevertheless, beyond the
two multidisciplinary journals leading the list, there are two physics
journals (Physical Review Letters and Astrophysical Journal) and one from
chemistry (Journal of the American Chemical Society) in the top 20. These
three journals – the most prestigious in their fields – ranked outside the
top 100 by impact factor, which demonstrates the slightly more balanced
character of this indicator.

On the other hand, the highest journal h-index in mathematics is 12, for
the Journal of Functional Analysis, which, with a multiple tie somewhere
around the 1500th position is certainly meaningless if the real "impact" of
the journal is sought. Hirsch's h-type indices will certainly challenge
scientometrists and other number crunchers for a while.

Tibor Braun
Wolfgang Glänzel
András Schubert
Hungarian Academy of Sciences
Budapest Loránd Eötvös University
Budapest
Steunpunt O&O Statistieken
K. U. Leuven, Leuven
Hungary.
braun at mail.iif.hu