In 2012, this journal admitted its gender bias. Following a complaint from two readers that too few News & Views articles were written by women, we totted up the numbers and realized that they were correct. Moreover, the imbalance was present in other sections of Nature, too. At the time, we pledged to commission more female scientists as reviewers and writers by asking editors to explicitly consider them, and to report back on progress (Nature 491, 495; 2012). We did so in 2013 and the results were mixed. There was progress, but it was patchy and we conceded that we needed to keep trying, and to try harder (Nature 504, 188; 2013).

It is time for another update, not least because the issue of gender imbalance in scientific publishing is the subject of a Comment piece this week. The authors analyse data from the American Geophysical Union (AGU) and find that female scientists are under-represented as reviewers of academic papers in the organization’s journals. They also show that female and (especially) male authors and editors recommend too few women as reviewers. (Such recommendations are a common feature of peer review, but editors are under no obligation to follow them.)

Between 2012 and 2015, the latest analysis shows, 20% of reviewers of papers in the AGU journals were women. This contrasts with the 28% female AGU membership and the 27% of female first authors on AGU papers during the same period. The Comment authors say that this gender bias is mainly because of the disparity in the number of female reviewers suggested.

Internal data on the gender of reviewers of Nature papers show a better picture — but not by much. In 2014, 23% of our reviewers were women, and 22% in 2015. Those figures are lower than we would like, but they do show a marked improvement on previous years. In 2011, just 14% of Nature reviewers were women, with 12% in 2012 and 13% in 2013. (Figures for 2016 have not yet been compiled — a non-trivial task that involves the manual looking-up of names to attribute gender.)

Nature does not routinely gather and collate data on the gender of recommended reviewers. But we asked our manuscript editors to perform an informal survey of those suggested for ten recent papers they have handled. Just 141 of 1,157 recommended reviewers (12%) were women.

There are some reasons for this low figure, and these should be considered in any robust analysis of gender bias in academia. More physical-science editors responded, and those disciplines have fewer women in senior positions. Indeed, the whole issue of seniority (the participation of women in science tails off as researchers climb the career ladder, for well-explored reasons) might help to explain why so few reviewers recommended to Nature by male and female authors are women. Interestingly, the Comment authors report that the bias they identified was present at all ages, and suggest that the lack of female reviewers is not just down to the seniority issue.

Female participation continues to grow, although not across the board.

At Nature, we often see the most senior of researchers in reviewer recommendations, but authors are asked to recommend rising stars too. This pool of potential referees has a much greater proportion of women. European Union statistics for some relevant fields show that women hold 33% of assistant professor and 24% of associate professor positions, compared with 13% of full professorial roles. We often pair a new referee with one who is tried and trusted to see how the new referee fares, to enable them to see reports from more-experienced reviewers and to give them an insight into our decision-making processes.

In the magazine section of Nature, female participation largely continues to grow, although not across the board. Some 25% of News & Views authors in 2015 and 2016 were women (up from 12% in 2011 and 19% in 2013). Women wrote 23% of World View articles in that combined period (12% in 2013) and 20% of Comment pieces in 2016 had a woman as a first author (27% in 2013). In News Features, 56% of full researcher profiles were of women in 2015, compared with 66% in 2016 (in 2011 and 2013, the proportions were 18% and 40%, respectively).

We hope that readers find these formal and informal statistics thought-provoking. We encourage other scientific organizations to track their own rates of gender participation, and we will continue to try to improve and report on our own.