GENDER BIAS IN PEER REVIEW AND SCHOLARLY PUBLISHING

It has been challenging to assess gender bias across scholarly publishing in part because data on both gender and age of participation are needed, as the proportion of women authors and reviewers varies greatly with age across most disciplines. The American Geophysical Union (AGU) membership is 29.3% female (from 2013-2015), and broadly representative of the gender distribution of employed Earth and Space Scientists in the U.S. Using AGU’s demographic data from its membership database, we matched individuals participating in AGU’s peer review and publication process from 2012-2015 to produce a large dataset (about 100,000 author and reviewing transactions) of gender and age of authors and reviewers for AGU journals to examine implicit bias in peer reviewed scholarly publishing.

Although women had higher acceptance rates than men (60.4 vs. 56.4%), they submit papers, especially as first authors, less often than their male peers. Female first authors submitted an average of 1.8 manuscripts each, while males submitted 2.1 each from 2012-2015. Women were also used less often as reviewers (17.9% of the time) than expected based on their membership in the society and their rate as distinct accepted first authors (26.7% female) and all distinct accepted authors (23.3%). This difference is apparent regardless of age, but greatest in the youngest cohort of scientists. Additionally, female reviewers each did an average of 3.65 reviews from 2012-2015, which is less than the 4.34 reviews per male reviewer. The bias is a result of fewer suggestions of women reviewers especially by male authors and editors, and also a higher decline rate by women within each age cohort when asked to review. Participating as a reviewer can have important career benefits, especially for early career scientists; thus, addressing this bias is important for addressing pipeline issues and improving retention of women in the field. Comparison with other studies suggests that these results are general.