The Best Mentors Are Creative And Encourage Independent Thinking, A New Study Suggests

We know a good mentor when we see one, but what exactly makes a great mentor? There is no one single way to mentor, and oftentimes, it’s a rather abstract and unintuitive set of qualities that coalesce to make a good mentor.

​

In one of the largest ever studies on mentorship, researchers tried to break things down and see what makes a good mentor. According to their findings, it’s the tacit, unspoken communication that makes the difference.

In the study, researchers look at protégé performance over the course of a career, drawing from genealogical datasets that track the relationships between mentors and students in science. Specifically, they looked at data from 40,000 scientists, assessing their scientific performance. Protégé success was determined by whether they won a scientific prize of their own during their career, were elected to the National Academy of Sciences or were in the top 25% of citations for their field.

They then compared an official record of advisor/student relationships taken from Ph.D. theses, and supplemented it with additional crowdsourced data from AcademicTree.org and the Mathematics Genealogy Project.

To compensate for the fact that successful mentors at big universities tend to attract more competitive students, the research group also grouped mentors with similar records and reputations, grouping them based on their institutional resources and productivity. They only compared mentors with similar metrics. Simply put, they tried to make for a fair comparison and see what difference the mentorship style itself made.

Of course, it’s hard to quantify how successful a career was and everything that goes into mentorship, but it’s as good a comparison as it gets — and the study revealed remarkable patterns.

In all the mentor groups, regardless of resources and prestige, some mentors were consistently capable of identifying and solving key problems. The most successful protégés studied under mentors that had unique ideas — but it gets even more interesting. The protégés working under these mentors also tended to have the most success when they strayed from the research direction of their mentors.

This suggests that the mentors that make unique contributions and encourage their students to think independently have the biggest impact.

Researchers also found that when mentors excel in transferring tacit knowledge, their protégés achieve two to four times greater success than similarly talented students of mentors who convey regimented knowledge but not tacit know-how.

“Communicating codified knowledge is relatively straightforward,” said corresponding author Brian Uzzi. “It’s written down in books and presentations. But it’s the unwritten knowledge we intuitively convey through our interactions and demonstrations with students that makes a real difference for mentees.”

The team also reports an important finding, especially in this day and age: face-to-face mentoring is much more important than remote communication.

“Face-to-face interaction is essential. When we teach by doing, we are conveying tacit knowledge we don’t even realize we have,” said Uzzi, the Richard L. Thomas Professor of Leadership at Northwestern’s Kellogg School of Management and co-director of the Northwestern Institute on Complex Systems. “If we limit the face-to-face channel by which tacit knowledge is communicated, we potentially slow down the pace of learning and scientific breakthroughs, and that will affect us all.”

This is one of the most comprehensive studies on this topic, and one of the very few to assess the lifetime achievements of both mentors and protégés. The world of academia sure has a lot of great mentors but, unfortunately, there are also quite a few negative examples. Studies like this can pave the way for healthier and more productive mentoring, which could make a world of a difference, both in academia and outside of it.

Journal Reference: Yifang Ma el al., “Mentorship and protégé success in science,” PNAS (2020). www.pnas.org/cgi/doi/10.1073/pnas.1915516117

Source