Since the 1950s, assessments of the quality and effectiveness of scientific research have increasingly rested on quantitative measures based on publication citations. The systematic analysis of academic citations – bibliometrics – has emerged as a new scientific field. Of course, citations can be biased – if scientists frequently cite their own works, for example – or if pairs of scientists cite one another to boost their citation numbers for mutual benefit. However, in a new paper, LML Fellow Fabio Caccioli, together with colleagues Weihua Li, Giacomo Livan and Tomaso Aste from University College London, note that few studies have considered how such reciprocating strategies might skew citation figures. Using data spanning more than a century, they’ve tried to fill this gap.
Caccioli and colleagues used the citation history of papers published in the Physical Review (PR) corpus of journals to find out, first, if the fraction of reciprocated citations can be used to predict scientific success and to classify different career trajectories, and, second, how prevalent citation reciprocation is in the scientific community at large. On the first question, they found a significant tendency for more successful scientists (as judged by citation counts) to be associated with lower levels of reciprocal citation behaviour. The least successful scientists, by contrast, showed the highest reciprocating behaviour.
The researchers also studied how reciprocating behaviour has evolved over time. Their results indicate the emergence over the past two decades of a clearly distinct “rich club” of top scientists who exchange citations at a rate much higher than would be predicted by a null network model of citations. Give the small number of such top scientists, however, most reciprocal citation in the entire network remains attributable to low to medium impact authors.
The paper is available at https://arxiv.org/pdf/1808.03781.pdf
LML is a charity. If you enjoyed reading this, please consider supporting us.