This past November, a heated conversation about mentorship in science was sparked by a Nature Communications study that alleged that female scientific trainees are less prolific in their subsequent careers if their research adviser is a woman. After serious backlash, the journal took a step back and retracted the paper.
The response highlights the positive impact of female scientists on female trainees, with many banding together to formally speak out against the journal and creating a database of mentoring pairs
numbering more than 1,500. As graduate student mental health reaches crisis levels, many graduate students cite poor mentorship as a large contributor.
On supporting science journalism
If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.
As an assistant professor of instruction at Northwestern University, I think about mentorship daily. Each year, I manage and train the incoming cohort of incoming Ph.D. students to serve as teaching assistants in our chemistry courses. Having received my Ph.D. in the last five years, I recall the stress and pressure that research demands and the power of a strong, compassionate mentor. My Ph.D. adviser, Bob McMahon at University of Wisconsin, understood that rigor and kindness are not at odds. I owe a great deal of my success to his guidance and understanding.
Perhaps instead of correlating success to selection of a mentor of a particular gender, it would be more productive to look at the structure of scientific mentorship. It is within this structure that scientists of all genders often fail their students, even with the best of intentions.
In 2019, the U.S. National Academies of Sciences, Engineering, and Medicine released a report arguing that U.S. Academic STEM mentoring does not, on average, follow best practices. Writers of the report stress that there is a science to effective mentoring, but that in most institutions, the mentoring structure is left up to chance. Scientific mentorship is wildly variable even within individual departments, and the culture of the research group is usually set by the adviser. Within academic institutions, there is no formal oversight for how professors manage their graduate students short of punitive action for explicit discrimination and harassment.
The bandwidth available for directed mentoring is often a function of the time an adviser has available. Ironically, the pool of interested mentees is largest when the adviser is at the top of their career, when their time is the most limited. As scientists gain prestige, they are more frequently called away to travel, leaving less time for direct contact with students. The pandemic has complicated the structure of mentor-mentee interaction, but it has also provided an opportunity to create more robust channels of communication. Through transitioning from primarily in-person meetings to Web conferencing like Zoom, mentees are no longer at the mercy of their mentor’s travel schedule.
The National Academies report, titled “The Science of Effective Mentorship in STEMM [the extra M is for Medicine],” also provides suggestions for more productive mentoring. There are nine major recommendations that center on transparency and accountability. Structured feedback systems between mentors and mentees, evidence-based methods for assessing quality mentorship, and integration of mentorship quality into tenure and promotion are all clearly outlined. Resources and plans of action are available; now it is in the hands of departments to execute them.
Using formal recommendations to improve the features of mentoring can go a long way toward improvement, but systemic change is often difficult to implement. The existing structure has persisted for a long time, and many believe that it is perfectly effective.
In 1985, Brian Martin and Jill Bowling described academic science as a patriarchal system that incentivizes masculinity in its very structure. Although this account is 35 years old, the underlying ethos that they describe persists. This ethos is often framed as objectivity, a way to ensure the integrity of the discipline, but it serves as a gatekeeper, maintaining the status quo.
This reveals an important truth about gender in science and how it manifests in mentor–mentee relationships. The identified gender of the mentor makes little difference when success requires shedding the type of empathic management style that is seen as “feminine.”
When discussing this issue in a course I teach for STEM graduate students and postdocs called “Equity in STEM for All Genders,” a Ph.D. student in engineering, who is a Black woman, let out a big sigh, stating “I don’t really care if there are more women scientists. They are all playing the same game.”
For those already fighting for a seat at the table, like female, trans or queer scientists, adapting to the utilitarian, “objective” view of group management often feels necessary for survival. In this way, hostile power dynamics are perpetuated regardless of gender. A great deal of effort has been directed to exhaustively tabulating the gender of science mentors, often using binary gender and inherently marginalizing trans and nonbinary scientists. Instead, science needs to develop leaders who prioritize healthy research group dynamics and are not penalized for the accompanying hit to productivity.
There is hope for change. As in many workplaces, COVID-19 has illuminated the inequitable burden female researchers have to bear. When research demands remain high, and the mentoring environment is not one that allows trainees to be heard and acknowledged, women’s careers will suffer. These hardships are bringing much-needed attention to the structural barriers that academic research places on developing scientists who don’t identify as male. Gender equity initiatives that have been advocated at the fringes for decades are getting press in major scientific publications.
Even major STEM corporations are recognizing injustices of the past and trying to make amends. After being fired by IBM in 1968 for coming out as transgender, Lynn Conway finally received a formal apology from the company in 2020. Across the gender spectrum, STEM professionals are finally being vindicated as the system reckons with its toxic history.
As for mentorship, institutions that want to make these changes can do so. A study of nearly 40,000 scientists who published close to 1.2 million papers in biomedicine, chemistry, math, or physics between 1960 and 2017 reports the far-reaching effect of strong mentorship on success of scientific protégés. The study indicates that fostering protégé success involves careful development of intellectual independence. Successful mentors are supportive and available, but they do not hold dominance over the careers of their trainees.
One effective model, advocated in “The Science of Effective Mentoring in STEMM,” involves several faculty members in the mentorship of each graduate student. Using a committee model allows students to feel like the entirety of their career is not in the hands of one individual, promotes intellectual autonomy, and allows access to guidance when their adviser is unavailable.
The Center for the Integration of Research, Teaching and Learning (CIRTL) Network offers an online research mentor training course for faculty and postdocs to improve research mentor skills. Other institutions like the University of Minnesota offers a similar program locally (in person or online).
Mentorship is an important component of any technical field, but in science it functions to both develop technically competent researchers and foster passion for discovery and grappling with the unknown.
This is an inherently emotional and engaging process, and a more humanized mode of mentorship is paramount.
Mentors must actively engage trainees while listening and allowing for their intellectual independence. In chaotic and unpredictable times, creating a sensitive and dynamic generation of scientists begins by example.