Page 175 - Becoming Metric Wise
P. 175
167
Journal Citation Analysis
calculate an impact factor for the field of mathematics, chemistry or
information and library science, as examples.
For a set of journals, J i , i 5 1, .. . n, one can define the global impact
factor (GIF) as:
n μ
P
i51 C i C
n μ
GIF 5 P 5 (6.9)
i51 P i P
where C i denotes the number of citations received by the i-th journal
(over a given citation window) and P i denotes the number of publications
in the i-th journal (during a given publication window). This is essentially
the same formula as that used for an impact factor of a journal. It is also
equal to the average number of citations per journal (μ C ) divided by the
average number of publications per journal (μ P ).
Yet, one might also calculate the average impact factor (AIF) of this
same set of journals:
n
1 X C i
AIF 5 (6.10)
n P i
i51
The difference between the two is essentially a matter of weighting
and hence of perspective. If one uses geometric means instead of arithme-
p ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
n
C 1 ?C n
tic ones then GIF(geometric) becomes p ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi while AIF(geometric)
n
P 1 ?P n
r ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
becomes n C 1 ? C n , which means that in their geometric form GIF and
P 1 P n
AIF coincide (Egghe & Rousseau, 1996b).
Normalization is further discussed in the context of indicators and
research evaluation.
6.7.4 Citable Publications
In the standard JIF 2 only the number of so-called citable articles is used
for the calculation of the denominator, while citations to all publications
in the journal are included in the numerator. Corrections, meeting
abstracts, book reviews, obituaries and short letters to the editor rarely
receive a large number of citations (Hu & Rousseau, 2013), and are for
this reason not included in the calculation of the denominator. An argu-
ment for this practice is that otherwise journals would be “punished” for
publishing these otherwise useful types of publications. It is, however, not