Recently the paper, “The association between early career informal mentorship in academic collaborations and junior author performance”, was published in Nature Communications. The authors are all professors in Social Science or Computer Science, the journal is highly respected, and the problem the authors try to address is indeed paramount: the role of mentorship in academia. The study puts forth a few interesting findings that I encourage people to read in full, but in particular, a main finding “suggests that female protégés who remain in academia reap more benefits when mentored by males rather than equally-impactful females” and that “increasing the proportion of female mentors is associated not only with a reduction in post-mentorship impact of female protégés, but also a reduction in the gain of female mentors.” (It is important to note that the study actually studies the relationship between gender and co-authorship, which was raised by the reviewers (See full Peer Review file for more details on this and other concerns)).
When I first read this headline, I did a cheeky eye-roll because I do not believe this finding to be accurate and is instead a result of a multitude of unaccounted confounding factors in the analysis. For example, the dataset used in the study comprises of publications from over 100 years ago and back to the 1800s. Well, WOMEN COULDN’T EVEN VOTE OVER 100 YEARS AGO, so do you think they were mentoring their own research groups? Although I doubt the 19th amendment has a direct implication in the study, I do think that deep societal and historical pressures could potentially explain this observed correlation between female mentors and protégés. I don’t want to do a full scientific breakdown of potential study design flaws, but it’s important to see that effect sizes and contingency tables provide evidence, but the story that’s crafted from this evidence is not always simply an objective cause and effect.
Later in the evening, said cheeky eye-roll turned into a longer thought process. I think I didn’t want to face what this article actually told me, what years of fighting imposter syndrome told me not to believe: my gender will always affect my career and that an academic workplace won’t welcome me as much as my male counterparts. From a current mentee perspective, I vehemently reject the notion that female mentors will be a detriment to my career. On a personal level, this point is moot. But now this article has planted the seed for so many others. I can already see the sound-bite reaching mainstream media and departments citing it in an argument to end funding for diversity-related mentorship. No one will likely say out loud, “You don’t want to work with a female PI.” (or maybe they will…the world continuously surprises me), but given the opportunity to rotate with a female or male PI of equal caliber, would someone not refer back to said sound-bite, despite it being perhaps a corrupted sound-bite?
One of my personal draws to academia has been the aspect of mentorship, where I could be the mentor as many others have been and currently are to me. But I sometimes think, will anyone want me to mentor them because of my gender? When I was a TA, it was clear that students preferred asking male TAs questions over myself and the other female TA. Since discussion sections have optional attendance, I wondered if students would even come to my section when another section is taught by a male TA at the same time? When I mentor an undergraduate student, am I who they were expecting to see on their first day? These are the types of questions that already plague me, and I’ve only been in the game for a few years.
I think the worst implication of this study is that it could very likely be repeated with a variety of minority groups, further shading academia as an unwelcoming environment for these members. Similar societal and historical barriers that prevent women from “post-mentorship impact” (e.g. many highly cited papers as per this article’s definition) similarly affect many racial minority groups. The authors then advocate for policy makers to consider the “consequences” of diversity policies with respect to “maximizing their long-term scientific impact”. Instead of advocating for future policy change in spite of diversity, I implore readers to instead think about what systematic barriers currently exist in academia that cause the need for such diversity policies. As much as researchers want to believe that they perform research for the pure sake of scientific discovery, their scientific findings have weight in the real-world, outside of their carefully controlled covariate studded studies.