Papers with female first authors receive 10% fewer citations than comparable papers with male first authors, a new study in Nature Astronomy reports, with potential implications for women’s career trajectories. “Citations from publications are currency—that is your net worth in academia,” says Anna Kaatz, director of computational sciences at the Center for Women’s Health Research at the University of Wisconsin in Madison, who was not involved in the work. “The less you’re cited, the lower your h-index, the lower your [Relative Citation Ratio], the lower status you are as a scientist—regardless of how meritorious your work is.” As lead author and Ph.D. student at the Swiss Federal Institute of Technology in Zurich Neven Caplar puts it, “If I have 10% more citations than a woman with the same quality paper, I would be more likely to get the job.”
The study looked at nearly 150,000 papers published in five major astronomy journals between 1950 and 2015. The citation gap has closed considerably since the earlier years in the data set, when men received 50% to 100% more citations than women. But since 1985, when the difference started leveling out, papers from men still received 6% more citations than papers from women. After normalizing for the papers’ features—including journal, publication year, subdiscipline, seniority of the first author, number of authors, and number of references—those with female first authors still received just 90% of the citations that would be expected if the lead author was a man.
Based on this analysis, gender bias—not paper quality—appears to be the most likely explanation for the citation gap, Caplar notes, though factors such as tendency to self-cite, collaboration networks, and conference attendance would also be worth exploring. Scientists choose to cite the papers that have the most value to them, and “whenever it’s a subjective judgment, I feel that women lose,” says Meg Urry, professor of physics and astronomy at Yale University.
Studies like this one can come across as yet another example in the seemingly endlessly expanding body of evidence that women are disadvantaged in science. “It’s easy to say ‘Oh my god, here we go again,’” acknowledges Urry, who has been working to increase the participation of women in astronomy for the past 30 years. Cases come up throughout the pipeline, Kaatz says. For instance, there is evidence that women receive lower quality mentoring, and that female faculty members tend to be assigned service and teaching responsibilities that can take away from their time spent on research. Other recent examples include reports that female first-year biology Ph.D. students are less likely to publish, and that female geoscientists are less likely to be described as excelling beyond other students in reference letters. When trying to drill down to the root cause of these observed biases, “it's hard to pinpoint one thing,” Kaatz says. It’s probably a combination of factors that accumulate over time. “I don’t think you can uncouple them.”
But, Urry emphasizes, “the thing we have to do is remain vigilant about the data, and when we see a study like this one that says things aren’t being done fairly, we really have to take it seriously.” Notes Jevin West, assistant professor at the University of Washington’s Information School in Seattle, “you want this preponderance of evidence [so that] there’s not a lot of question of whether or not it’s true.” West is also excited that the study brings a new approach—a computational technique that has been shown to be good at predicting changes such as in housing prices—to analyzing data for gender biases. He is looking forward to applying it in his own work on the topic.
Urry is also enthusiastic about the possibility that the study’s methods could be used to create a tool that authors could use to check for biases in what they cite in their own papers. “Wouldn’t that be great?” she asks. When she writes recommendation letters, she uses such a tool—based on findings from a 2007 study—that flags male- and female-associated words to help her make sure that her letters don’t fall into the traps of gender-coded language. “The power of these analyses is tremendous,” she says.