“A woman needs a man like a fish needs a bicycle.”
– Irina Dunn
As medical journals scramble to endorse equity, diversity and inclusion (EDI), it’s clear that certain editorial boards have held onto questionable baggage.
On Nov. 17, Nature Communications saw fit to miscommunicate the importance and value of non-male mentors, a case of questionable peer review that rivals the #MedBikini misstep courtesy of the Journal of Vascular Surgery. While the hapless authors of that latter piece waded into a non-debate about two-piece bathing suits and medical professionalism, Nature’s brainchild, “The association between early career informal mentorship in academic collaborations and junior author performance,” takes things one step further and suggests that associating with females in any guise is dangerous.
To paraphrase Scrooge, are there no editors? Two of the three authors are women. All of the authors are affiliated with New York University Abu Dhabi, in the United Arab Emirates. Their expertise is largely in computational sciences and they espouse what they describe as the “field of Science of Science,” which is clearly distinct from any philosophy of science.
Their first and lesser value-laden assumption relates to the definition of mentorship. They assume that co-authorship implies mentorship, an assumption questionably supported by a survey with a response rate of less than 10 per cent in which those responses endorsed some definition of mentorship barely 50 per cent of the time. (Their graphic is captioned “Responses of 167 randomly chosen scientists” when it ought to read “167 responses from 2,000 randomly chosen scientists”).
Further, the authors are explicit about publication quantity determining the strength of a mentor; they describe “big shot” mentors as those who have accrued a large “average number of citations per annum up to the year of their first publication with the protégé.”
Mentorship, however, exists on a spectrum; one can get career advice from colleagues who co-author publications but the depth of a mentor-protégé relationship transcends and is certainly not proportionate to publication output – the more generically mentorship is defined, the more one will find an association with broader existing trends in the scientific community such as the barriers to non-male, non-white scientists.
The authors mined a large Microsoft database of scientific publications and arrived at the rather unsurprising conclusion that women are not afforded equal opportunity in the medical sciences – they report that female mentors seem to suffer (in terms of publication output) from mentoring females and female protégés suffer from being mentored by females. Regardless of the flawed definition of mentorship, the flawed notions of success and the flaws of drawing such far-reaching conclusions from large datasets, the associated conclusion – that male scientists are afforded a disproportionate advantage – is unfortunately likely to be more true than false. They could have stopped there and New York University Abu Dhabi might even have used this data to publish a scathing indictment of the medical research establishment.
Instead, the authors decided that the “is” abstracted from their data dictates an “ought” and concluded that “current diversity policies promoting female-female mentorships, as well-intended as they may be, could hinder the careers of women… female scientists, in fact, may benefit from opposite-gender mentorships in terms of their publication potential and impact throughout their post-mentorship careers.”
Outcry ensued: “This paper should not have been published. The raw scientific results may be of value but the interpretation of them is extremely flawed” and “I am ashamed as a human to read a pseudoscience article like that published in an accredited journal. When I read the news about this article, I thought it was a hoax” were among some of the online comments.
The paper eventually may go the way of the retracted #Medbikini piece but retractions do not solve the systemic and oft-unspoken inequities made explicit in publications such as these; and the outcry must be levied against more than a single publication.
Overall, this paper’s saving grace is its timely illustration of why equity, diversity and inclusion face a particularly uphill battle in medicine.
First, the paper’s reductionist view of mentorship-as-publication-quantity is emblematic of the typical short-sighted metrics of success in medicine that disadvantage those who give birth to a child or those who pursue essential roles in teaching, advocacy and clinical practice that do not leave time for rampant randomized clinical trials. Inclusion is not simply about throwing doors open to those few who can squeeze into some outdated measure of success; until medicine can build a holistic way to define, support and celebrate success for its clinicians, administrators, educators and researchers, it will always face a fundamental conflict of interest in its endorsement of EDI.
Second, equity cannot be achieved if something as fundamental to personal identity as gender is treated as if it were some variable that ought to be modified, judged, valued or manipulated (rather than understood in some way, e.g., the many studies which have linked race to poor COVID-19 outcomes). To treat persons and their inherent qualities as associations to be avoided is unethical (i.e., judging some facet of personhood because of an outcome, rather than judging an outcome because of its assault on personhood). I don’t believe that was the intention of the authors but they made the mistake of letting the data do the speculating.
The valuable mentorship afforded by female scientists is literally self-evident (i.e., predicated on the self) and the self is never a means to an end. Scientific organizations should loudly endorse the importance of gender equality in mentorship opportunities and this endorsement should be robust without seeking evidence that such mentorship produces some threshold number of Nature Communications publications.
Indeed, this paper highlights the general pitfalls awaiting anyone who is eager to embrace diversity as an evidence-based science. Evidence of non-diversity is welcome, adding to the thankfully growing clamour. But evidence for diversity (i.e., for its effectiveness in achieving X, Y or Z) can never be allowed to become a necessary starting point for what must ultimately be a values-based exorcism of the systems opposed to that diversity.
The data, if left to speak for themselves, will only speak power to truth.