Publication: Quantitative evaluation of alternative field normalization procedures
Loading...
Identifiers
Publication date
2013-03
Defense date
Advisors
Tutors
Journal Title
Journal ISSN
Volume Title
Publisher
Impact
Export
Abstract
ide differences in publication and citation practices makes impossible the direct comparison of raw
citation counts across scientific disciplines. Recent research has studied new and traditional
normalization procedures aimed at suppressing as much as possible these disproportions in citation
numbers among scientific domains. Using the recently introduced IDCP (Inequality due to Differences
in Citation Practices) method, this paper rigorously tests the performance of six cited-side
normalization procedures based on the Thomson Reuters classification system consisting of 172 subfields.
We use six yearly datasets from 1980 to 2004, with widely varying citation windows from the
publication year to May 2011. The main findings are the following three. Firstly, as observed in
previous research, within each year the shapes of sub-field citation distributions are strikingly similar.
This paves the way for several normalization procedures to perform reasonably well in reducing the
effect on citation inequality of differences in citation practices. Secondly, independently of the year of
publication and the length of the citation window, the effect of such differences represents about 13%
of total citation inequality. Thirdly, a recently introduced two-parameter normalization scheme
outperforms the other normalization procedures over the entire period, reducing citation
disproportions to a level very close to the minimum achievable given the data and the classification
system. However, the traditional procedure of using sub-field mean citations as normalization factors
yields also good results.