Saturday, January 09, 2021

An analysis of 145 journals suggests peer review itself may not explain gender discrepancies in publication rates

Peer review and gender bias: A study on 145 scholarly journals

AMERICAN ASSOCIATION FOR THE ADVANCEMENT OF SCIENCE












Research News

An analysis of 145 scholarly journals found that, among various factors that could contribute to gender bias and lesser representation of women in science, the peer review process itself is unlikely to be the primary cause of publishing inequalities. However, Flaminio Squazzoni and colleagues emphasize that the study does not account for many other factors that may affect women's representation in academia, including educational stereotypes and academic choices of priorities and specialties. Even as female representation has improved in fields such as the humanities, psychology, and the social sciences, a publication gap persists, with male authors continuing to publish more manuscripts in more prestigious journals. To better understand whether peer review and editorial processes contribute to these gender discrepancies, Squazzoni et al. leveraged an agreement on data sharing with several large scholarly publishers; the team includes authors from the publishing companies Elsevier, John Wiley & Sons, and Springer Nature. The researchers collected and analyzed almost 350,000 submissions to 145 journals by about 1.7 million authors, as well as more than 760,000 reviews completed by about 740,000 referees. The sampled journals were identified by publishers so as to maximize coverage of research fields, although journals from learned societies, among others, were not considered. They then analyzed each step of the editorial process for bias, including the selection of referees, referee recommendations, and the editorial decision for each manuscript. They also accounted for each submission's research field, its proportion of women authors, and the position of women in the author list, while controlling for the proportion of women among referees, journal impact factors, the number of authors per manuscript, and the type of peer review adopted by each journal (single-blind or double-blind). The authors note, however, that it was not possible to directly estimate the quality of each submission. Author gender did not appear to affect how frequently manuscripts were accepted in the life sciences and social sciences, while manuscripts with higher proportions of female authors were in fact more likely to succeed in biomedicine, health, and physical sciences. "Our findings do not mean that peer review and journals are free from biases," the authors write. "For instance, the reputation of certain authors and the institutional prestige of their academic affiliation, not to mention authors' ethnicity or the type of research submitted, could influence the process, and these factors could also have gender implications." The researchers note that collaborative data sharing efforts from funding agencies, academic institutions, and scholarly citation databases will be necessary to further elucidate how existing structures determine academic opportunities.

No comments: