Page 279 - Becoming Metric Wise

271
Research Evaluation

• Sequence of CpP in the designated database (sequence v and v ). The
0
values of sequences t and t divided by the corresponding number of
0
publications in the designated database (differing by the year).
• Weighted (by number of publications) average diachronous impact fac-
tor of journals used by the research group calculated over the same
citation window (denoted as JCSm).
• World citation average in the field (denoted as FCSm) (with or with-
out self-citations).
One of the weaknesses of a designated database such as the Journal
Citation Reports for evaluation purposes as well as for studies of the
science of science, is the fact that many journals are not in the correct
category. Moreover, there is no classification of articles, in the sense
that articles are assigned to the category of the journals in which
they are published. Consequently there is a need for a better, article-
based, classification scheme, see e.g. (Waltman & van Eck, 2012b). In
an attempt to produce such a scheme, Gla ¨nzel and Schubert (2003)
note that the classification of the scientific literature into appropriate
subject fields is one of the basic conditions for valid scientometric
analysis.
• Ratio of average impact of the group and average impact of the
journal package CpP/JCSm (or CexpP/JCSm).
CexpP is CpP but with research group self-citations removed.
JCSm is the journal citation average.
In older publications one finds the so-called crown indicator,
defined as the ratio of the impact of the research group and the world
citation average CpP/FCSm. This indicator became known as the
Leiden crown indicator, but was actually used earlier by the research
group of the Hungarian Academy (Braun-Schubert-Gla ¨nzel). Because
of problems with this indicator (it has namely been shown that it is
not independent, see Subsection 7.3.3), it has been replaced by the
mean normalized citation score (MNCS). This indicator can roughly
be described as the average number of citations of the publications
under investigation, normalized for field differences, publication year
and differences between document types. An MNCS value 1 repre-
sents the world average. Hence, an MNCS value of 2, for instance,
means that these publications have in total been cited twice above
their field’s world average.
• JCSm/FCSm.