Page 165 - Becoming Metric Wise

157
Journal Citation Analysis

Table 6.2 The matrix needed to determine an impact factor (derived from Table 6.1)
Year of publication 1 2 3 4

Number of articles 2 3 3 4
Number of citations received in year 1 1
Number of citations received in year 2 6 4
Number of citations received in year 3 8 11 3
Number of citations received in year 4 9 14 6 10

The purpose of Table 6.2 is to show that it can easily be derived from
the complete p-c matrix (Frandsen & Rousseau, 2005; Ingwersen et al.,
2001). Information about individual cited or citing articles is not neces-
sary. Only the number of publications (per year) and the total number of
received citations per year by all publications matter.
The first row of Table 6.2 gives the publication year, while the sec-
ond one shows the number of published articles in each of these years.
For simplicity we assume that each article is taken into account when
calculating a JIF. In reality, only some types of articles are. These are
referred to as being “citable.” Editorials, meeting abstracts, book
reviews and similar publications are considered “uncitable.” We go into
some more details about this in Section 6.13. The following rows in
Table 6.2 are citation rows. The table shows, for instance, that in year 3
this journal received 8 citations for articles published in year 1.
Moreover, this journal received in year 3, 3 citations to articles pub-
lished in the same year.
Starting from such a p-c matrix a plethora of impact factors and even
types of impact factors can be defined. An important difference is that
between synchronous and diachronous impact factors (Ingwersen et al.,
2000, 2001). The standard Garfield-Sher impact factor for journals,
denoted as JIF 2 , is a synchronous impact factor based on one citation year
and two publication years (Garfield & Sher, 1963). An example: the
Garfield-Sher impact factor of the journal represented in Table 6.2 in the
year 4 is obtained as:
14 1 6
JIF ð4Þ 5 3:33
2
3 1 3
It is this impact factor which is published on a yearly basis in
Thomson Reuters’ (now Clarivate Analytics) JCR, a statistical byproduct
of the WoS, see further. A precise definition follows in Section 6.4.