Page 124 - Becoming Metric Wise

115
Publication and Citation Analysis

• Which database should one use: the WoS or Scopus? Or maybe GS?
Is it advisable to use a local database or a local database combined
with an international one? Recall that none of the international data-
bases is complete (and even local ones are not). Because the coverage
of databases tends to vary with disciplines, this reinforces the previous
point: the number of citations in a poorly covered field may in reality
be much higher than what one can deduce from one single database.
• Dominance of English in the sciences. English is the lingua franca of
the sciences. As a consequence papers published in English are pre-
ferred for citations (Liang et al., 2013). This means that if two papers
make the same point and one of them is written in English and the
other one in another language, the one written in English will receive
the most citations.
• Bias in favor of the United States, bias at the expense of developing
countries. In the eighties Cronin (1981) found that US authors cited
American works for 95% of the time, a percentage which is much
larger than the US’ share of publications. We do not know if this is
still the case, but expect that it is not. Bias at the expense of develop-
ing countries is related to the Matthew effect on the level of countries
(Bonitz, 1997; Bonitz et al., 1997). Note though, that almost all coun-
tries exhibit an own-country preference, in the sense that they cite
articles coauthored by countrymen relatively more than expected from
the share of the own country’s publications.
• Gender bias. Ferber (1986) claims that researchers tend to cite a larger
proportion of authors of their own sex than of the opposite one. This
has substantial consequences in fields where men are the large major-
ity. A numerical calculation illustrating this fact can be found in
Rousseau (2006a). Gender issues in scientific evaluation are briefly
discussed in Subsection 8.13.5.
• Papers containing important ideas will not necessarily continue to be
highly cited. Once an idea is widely known in its field citing the orig-
inal version is deemed unnecessary, e.g., Einstein’s theory of special
relativity. This phenomenon (no longer citing the original source) has
been termed obliteration by incorporation. It has recently been studied in
McCain (2012).
• Errors in databases. Citation analysis cannot be more accurate than the
raw material, usually the databases, they use. In turn, these databases
cannot correct every mistake made by writers when compiling