Image: Shutterstock

Machine Learning Reveals Systematic Sexism in Astronomy

Becky Ferreira

Becky Ferreira

Male astronomers are cited 10 percent more frequently than female astronomers in leading journals.

Image: Shutterstock

Sexism in astronomy has become a topical issue in recent years, especially after prominent UC Berkeley professor Geoffrey Marcy was forced to resign in 2015 over sexual harassment charges from many of his female students, which led to an outpouring of related stories. Discrimination against female astronomers has been shown to impact everything from equal pay to access to leadership roles.

Now, a quantitative study published on Friday in Nature Astronomy demonstrates that gender bias in astronomical research extends even to journal citations, which are an indicator of academic prestige and are linked with better access to grant money, speaking engagements, and professional advancement.

Led by Neven Caplar, a PhD student at ETH Zürich's Institute of Astronomy, the new research found that papers with male lead authors were cited 10 percent more frequently than papers led by women, even after controlling for non-gender-specific disparities such as seniority, team size, publication date, field, and academic institution.

This exposes "clear indications of the existence of gender bias in astronomy," said Caplar and his colleagues.

The team reached this conclusion after using machine-learning to analyze a dataset of over 200,000 papers published between 1950 and 2015 in five influential journals: Astronomy & Astrophysics, The Astrophysical Journal, Monthly Notices of the Royal Astronomical Society, Nature, and Science.

The team identified the genders of first authors by running bibliographic information through databases such as SexMachine or Gender API. They scrubbed any research in which the lead author's gender was inconclusive, reducing the final sample to 149,741 papers. In cases where first authors used their initials—a tactic women researchers disproportionately use to avoid gender bias—Caplar's team took extra measures to identify exceptions in publishing records that exposed authors' full names.

Read More: Bill Nye Didn't 'Censor' Gender Science, He Updated it Because That's How Science Works

"Even if you [normally] use your initials, but you used your full name once in your career, we could find you and match you by your surname and the first letter of your initials," Caplar told me over Skype. Women who changed their maiden names could also be identified this way. However, Caplar added that initialled female authors with only a few papers under their belts would be trickier to track down.

After the papers had been split by gender, the team cross-examined the various properties of male-led research, such as research field, paper length, and professional experience. They matched these non-gender-specific qualities with the female-led research and predicted the citation rate for female first authors if there were no gender bias, and compared it to the actual results shown in the sample.

Graph plotting female citations in astronomy journals over time. Image: Neven Caplar/Sandro Tacchella/Simon Birrer

As visualized above, women have been closing the citation gap steadily over the past half-century, but they still remain underrepresented within this metric.

The study also reveals a range of gender disparities across astronomical specialities. "We divided the subfields into six categories, and you can clearly see that men are more prominent in instruments and methods papers," Caplar told me. "On the other hand, you can see a strong increase in the number of women writers in the planetary science papers, because planets are a hot topic, and there are more women in astronomy now than before."

The fact that female astronomers are systematically cited less often than men likely has negative ricochet effects for their careers. When combined with previous studies reporting wage gaps of 25 to 40 percent between male and female astronomers, a tendency for female astronomers to have fewer children than they'd like, and the underrepresentation of women in publicly valued roles, such as astronauts or Nobel Prize recipients, it's clear there's much work still to be done to even the astronomical playing field.

This bias is not only a burden for women trying to get ahead in astronomy or another scientific field. As Caplar and his co-authors point out, "a pool of researchers from a wide range of backgrounds, experiences and perspectives maximizes creativity and innovation" in science.

When equally qualified women are unequally represented, the astronomy community as a whole suffers, as does our understanding and exploration of worlds beyond our own.

Subscribe to Science Solved It, Motherboard's new show about the greatest mysteries that were solved by science.