Women Scientists Are Calling Bullshit On a Study Claiming That Women Make Bad STEM Mentors

Illustration for article titled Women Scientists Are Calling Bullshit On a Study Claiming That Women Make Bad STEM Mentors
Graphic: Elena Scotti (Photos: Getty Images

Last week, the scientific journal Nature Communications, an offshoot of the internationally respected journal Nature, enraged scientists, women, and anyone who has ever been near a graduate program by publishing a study that concluded maybe women mentors are actually hurting the career prospects of their mentees. As the study’s authors wrote, the research “suggests that female protégés who remain in academia reap more benefits when mentored by males rather than equally-impactful females.”


Thousands of scientists from all over the world identified issues with the methods used in the study, titled, “The association between early career informal mentorship in academic collaborations and junior author performance.” Its methods and conclusions were even criticized by scientists who had peer-reviewed the paper, whose concerns were published alongside the study but were not addressed within the text. In response, the journal quickly promised an investigation of those methods and conclusions (the kind of thing that usually happens before publication, during peer review).

The immediate backtracking raised the question of why a respected scientific journal would publish seemingly unsound, sexist science at all, especially if the study’s own peer reviewers did not find its methods sound. The answer, according to several scientists, has to do with the shifting business model for scientific journals, which are increasingly incentivized to accept flashy, dramatic studies, which draw attention and readers but can sometimes turn out to be less accurate.

The paper, authored by two women and one man out of NYU Abu Dhabi, used an algorithm called Genderize in order to determine the genders of the authors and co-authors of 200 million scientific papers written over the course of 100 years, sourced from Microsoft Academic Graph. From there, the research segmented scientists into two groups: “junior” scientists, less than seven years into their careers, and “senior” scientists, more than seven years into their careers. According to Science, “They found that early-career scientists who co-wrote papers with what the authors call ‘big-shot’ researchers—defined by their yearly citation rate—went on themselves to have citation rates that were higher than average.” Furthermore, the data seemed to suggest that “the more female mentors an early-career scientist had, the lower the impact of the papers they published when they became senior scientists.” The study also reported that female mentors also “suffer on average a loss of 18% in citations on their mentored papers.’”

On Twitter, scientists both male and female came together to protest the conclusions drawn from the study—namely, not that the systems and structures of scientific institutions need to be overhauled to reach parity, but that junior scientists would likely benefit from working with male mentors. In a widely shared open letter to Nature Communications, Dr. Leslie Vosshall, neurobiologist at The Rockefeller University, wrote that those conclusions were based on “flawed assumptions and flawed analysis,” demanding a retraction of the paper. Others were quick to acknowledge the support of the women mentors responsible for encouraging their STEM careers. (Dr. Sarah Davies, a biologist at Boston University launched a Google Sheet in which over 1,500 scientists gathered to thank their mentors.)

Not only was the authors’ definition of mentorship limited—relegated to writing a paper with someone—the paper’s binary approach to gender, along with the inability of Genderize to always correctly predict gender based on name alone, muddles the conclusion that women mentors are inhibiting their mentees’ success. “They defined mentorship as co-authorship,” Dr. Rebecca Barnes, a biogeochemist at Colorado College and leader at 500 Women Scientists, a non-profit working toward gender parity in STEM fields, told Jezebel. “In science, you could author a paper with someone that you really have very little interaction with.” Furthermore, Barnes added, Genderize is known to have difficulties correctly identifying the gender associations with non-European names. In fact, the authors of the study noted that Genderize couldn’t classify nearly 48 percent of the names.

Mentorship in science is a complicated relationship, which includes not easily quantifiable forms of support and guidance. Though the authors of the paper claim to have identified 3 million mentor pairs, because of their limited definition, Barnes says that all the study actually did was identify 3 million co-authors. The study does find that male scientists do get more citations on average than their female counterparts, information that is not exactly groundbreaking to anyone who works in the field. Take a 2018 study published by Inside HigherEd (that admittedly also used Genderize), which found that men were much more likely to cite other men than women, even in journals with majority women authors. “We already know that science and publishing are biased in favor of men,” Dr. Barnes said.


While the study’s authors acknowledge that “there are social aspects that are not captured by our observational data,” the paper makes the interesting choice to instead focus on possible individual shortcomings of women mentors. “Female mentors,” they write “[may serve] on more committees, thereby reducing the time they are able to invest in their protégés.” They conclude that not only would women benefit more from opposite-gender mentor/mentee relationships but that mentors themselves might benefit more from working with male mentees.

According to Dr. Morgan Kelly, a biologist at Louisiana State University who rescinded her offer to peer review for Nature Communications in the wake of the study, this latest controversial publication is a symptom of a larger trend for the journal, reliance on novelty science to prop up a for-profit business model:

“My decision to not review for this journal anymore is proximally about this most recent paper, but it’s partially about the fact that I think this paper is about a larger problem with the publishing model,” Dr. Kelly told Jezebel.


Nature Communications is an offshoot of Springer Nature, a private-equity backed academic publisher that raked in over a billion dollars in revenue last year. In part, this funding comes from charging researchers eager to associate with the stalwart Nature brand a fee of over $5,000 in order to publish their studies in a splinter journal. This business model creates open-source information that can then be shared freely among the scientific community and beyond. But some scientists worry that the need for exciting content to draw readers to these journals combined with an abundance of researchers eager to be associated with the Nature brand also means a higher demand for science that earns attention, a phenomenon known as the “file-drawer effect,” which Inside HigherEd describes as a preference among scientific journals for research that has “‘positive’ and even flashy findings that support a hypothesis are more likely to be published than negative’ ones,” adding that the practice is “well known.”

But in this case, an attention-grabbing study could contribute to greater inequality within the scientific community. “There is a real and legitimate interest in studying inequality in the sciences,” Dr. Kelly said. “Trying to understand why women still make up the minority of tenured faculty in most academic departments, but I also think there was a troll element to this particular study.”


Dr. Kelly doubts that Nature Communications is going to have much effect on the number of students asking for her guidance, especially in academia where women do much more “service work” than their male counterparts, nor is she concerned fellow scientists will take any of the study’s conclusions too seriously. However, she is concerned that hiring committees will withhold resources from attempts to diversify after being given an indicator from a high profile academic journal that the effort isn’t worth it. This, she worries, will make life more difficult for the women and underrepresented groups who are currently working in academia:

“Universities are beginning to make progress in their efforts to have more diverse faculty,” Dr. Kelly said. “One of the reasons that I’m on lots of committees is because there are a little over 20 percent women faculty in my department, and quite often graduate students will ask me to be on their committees because they want at least one woman. That happens to women and minorities in the department because there are not that many of us.”


In a statement to Jezebel, Dr. Elisa De Ranieri, editor in chief of Nature Communications wrote, in part: “The concerns that were raised have implications for not only women but every single person in academia, and I’m committed to looking into the concerns around this paper carefully: not only the methodology and conclusions drawn, but also the editorial handling of the manuscript.” A note attached to the published research also notes that the company is “investigating concerns.” However, Nature Communications has yet to answer any questions as to why they did not investigate these concerns when peer reviewers first raised them.


Honey, where'd I leave the burner key again?

The sheer laziness of this study is breathtaking. Experienced male authors have higher citations and female researchers that co-author with the experienced males have higher citations. The established predominance of male researchers in STEM fields is just taken as a given, like the sun rising in the east and water is wet.

As a decades-long refugee from academia the Google sheet of researchers thanking their mentors doesn’t surprise me. The relationship between mentors and mentees has a significant transactional component, and a mentee pretty much has to ride or die, since going against one’s mentor (male or female) is academic suicide. Although there is definitely some variation, for the most part most mentors try to help their mentees, but their help is limited by the power/connections that they hold within the academic community. Again, any differentials that are observed between the perceived success under male versus female mentors reflects more on the underlying power structure than some differential desire to mentor.