Author gender and citation categorization: a study of high-impact medical journals

Abstract

The number of citations is one of the main bibliometric indicators. However, not all citations can be considered equivalent; scite (https://scite.ai/), a new tool based on artificial intelligence, was developed to determine whether citations are positive, negative or neutral. We assessed whether publications first/last authored by women were more often cited positively (or negatively) than those first/last authored by men. Using the 2021 Journal Citation Reports (JCR) impact factor, we selected the ten highest impact journals in nine medical disciplines. Using Web of Science, we extracted all research and review articles published between January 2012 and December 2021 in these journals. We used Namsor to determine first/last authors’ gender and scite to categorize article citations as positive (“supporting”), negative (“contradicting”), neutral (“mentioning”) and “unclassified”. There were 141,921 articles in the database, of which 116,204 had unabbreviated first/last names. We found that the proportion of positive and negative citations was higher for publications whose first/last authors were women (vs. men), while the opposite was true for neutral citations. This is the first study to our knowledge to document the association between gender and citation type. Further research is needed in the future to investigate the reasons for these gender differences, and to assess whether the type of citation is also associated with the gender of the citing author.