Women remain under-represented in many areas of science, but they are especially scarce in the pages of high-impact journals, according to an analysis published online 2 March in bioRxiv.

Researchers at the University of Washington in Seattle gathered names of first and last authors from papers published from 2005-2017 in 15 major science and neuroscience journals, including Nature, Science, PNAS, Nature Neuroscience and Neuropsychology Review. Nearly 10% of the names were excluded because they were relatively gender neutral, but the rest told a clear story: In these journals, authorship is a male-dominated enterprise.

For example, women accounted for roughly 25% of all first authors in Nature and Science and just over 35% of first authors in PNAS. Female first authors outnumbered men in only one journal, Neuropsychology Review, but just barely (53% vs 47%). Women made up an even smaller proportion of senior (or last) author spots, ranging from about 15% in Nature and Science to just under 40% in Neuropsychology Review.

The study found an inverse relationship between the prevalence of female authors and the impact factor of the journal—the higher the impact, the lower the chances that a woman was involved. Because publication in high-impact journals is so crucial for a scientific career, any gender gap could have serious consequences, says Ione Fine, a neuroscientist and co-author of the study. “If you aren’t published in high-impact journals, you don’t get awards or jobs,” she says. “It becomes a cascade of events.”

The scarcity of women in journals doesn’t simply reflect a lack of women doing high-quality science, Fine says. The study notes that roughly 30% of prestigious R01 grants from the US National Institutes of Health go to women. But in almost all of the journals studied, the percentage of women in senior author spots falls below that mark, a sign that the gender disparity in authorship exceeds disparities in other measures of academic excellence and productivity. “That’s the smoking gun that we have a real problem here,” she says.

Subtle biases by reviewers may make it harder for women to get published, Fine says. But she notes that women themselves may be contributing to the gender gap through a reluctance to submit to top-tier journals. “My feeling is that women are self-censoring because it’s just a more brutal process for them,” she says. “I know my male colleagues submit papers that I wouldn’t submit, and they seem to do just fine.”

Fine and colleagues call for all journals to keep statistics on papers submitted by women and minorities. They also suggest that journals could greatly reduce the possibility of bias by adopting mandatory double-blind reviews, a system in which the reviewer doesn’t know the identity—or the gender—of the study’s authors. Nature and other journals provide double-blind reviews on request, but Fine says that practice won’t protect women from bias. If an author requests double-blind review, she says, the reviewer is likely to assume that the request came from a female researcher, thus defeating the purpose.

In response, Nature Research, the parent organisation of Nature, issued a statement that read, in part: “Nature Research is committed to gender equality and our journals strive to support women in science.” The company says that it does not ask submitters to indicate gender, so it doesn’t systematically track gender statistics. It also says that it will “continue to assess the merits” of mandatory and voluntary double-blind reviews.

A 2017 Nature editorial noted that the journal has made slow progress in other areas of gender equality. For example, women accounted for just over 20% of reviewers in 2015, a small improvement over previous years. In 2013, 13% of reviewers were women. But Fine says that hiring more female reviewers won’t necessarily close the publication gap. “Women can be biased too,” she says.

Chris Woolston is a freelance writer in Billings, Montana.

Suggested reading:

Women in physical sciences

Fight the brain drain

Science is failing women