Citation rates are frequently used as a measure of research excellence, and while normalization procedures are usually applied to control for differences of discipline and year of publication (to create a level playing field for comparison), several other parameters are also known to have impact on citation scores: open access (OA) status, gender composition of research teams, and international co-authorship. These parameters are also known to correlate with each other; international collaborations have a greater likelihood of being available in OA, and women are known to participate less often than men in international collaboration.

The question addressed by this study, then, is how each of these three parameters influences citation rates. All three parameters are integrated into a multivariate regression model to distill the influence that each one has independently, controlling for the others. The study uses a custom dataset crossing data from Scopus (Elsevier) and the oaIndx (1science). The data on authors in Scopus was enriched using the NamSor API to determine their gender. The dataset is filtered to screen out confounding factors, focusing on one publication year, only those publications for which all authors can be unambiguously tagged as either women or men, and papers that are confirmed as peer reviewed in the production process of the oaIndx. Two subfields of research are considered—one dominated by male authors, and one closer to gender parity.

The study demonstrates that OA status has the largest magnitude of effect. Controlling for the number of authors and number of institutions involved in producing a publication, international co-authorship is shown to lend a citation advantage in one of the two subfields. Furthermore, the involvement of women in a research team is shown to have a positive impact on citation scores; however, in the subfield where overall authorship is closer to gender parity, the positive impact of women’s involvement diminishes as the share of women increases among the authorship group of a paper. These findings should inform the consideration of citation scores in research performance evaluation.

See the presentation here [PDF].

Image credit: iStock Photo