New scientists, better mentors?

How many graduate students a scientist advises over the course of his or her career and when he or she does so can affect how successful those students are in their own careers, and not in the way one might expect, according to a study published this week in Nature that looked at the field of mathematics.

Image: Wikimedia commons

Specifically, the researchers found that students appeared to fare well (measured by whether they eventually mentored many students of their own) if they were advised by mentors who trained fewer protégés, or by mentors who trained more students but were early in their careers. "I thought the paper was really provocative and interesting," said neuropharmacologist linkurl:Joan Lakoski,;http://www.pharmacology.us/Faculty.aspx?FacultyID=61 associate vice chancellor for academic career development at the University of Pittsburgh Schools of the Health Sciences. However, it remains uncertain "whether or not this study in mathematics translates to mentorship in the life sciences," added Lakoski, who was not involved in the study. Mentoring is a crucial part of science, but little is known about the effect of various mentorship styles on the protégés who grow up to be mentors themselves. A commonly cited hypothesis is that protégés emulate the style of their own advisors, with good mentors begetting good mentors -- the so-called "rising star hypothesis." But the evidence for such copycat mentoring is largely anecdotal, said social scientist linkurl:Edward O'Neil;http://nurseweb.ucsf.edu/www/ffoneie.htm of the University of California, San Francisco, who did not participate in the reserach. Interested to see if the rising star hypothesis would hold up to more rigorous analysis, statistical physicist linkurl:Luis Amaral;http://amaral.northwestern.edu/ of Northwestern University in Illinois and his colleagues examined the data from more than 7,000 mathematicians who graduated between 1900 and 1960. The data -- which included how many graduate students each mentor had advised (mentorship fecundity) and how many students those protégés had gone on to advise (protégé fecundity) -- were extracted from the Mathematics Genealogy Project, which has been collecting such information since the 17th century. The researchers found that mentors who advised more than the average number of students were more likely to be members of the National Academy of Sciences and have more publications, signs they were successful in their careers. But at first glance, mentorship fecundity seemed to have no correlation with the fecundity of their protégés. One exception, however, was individuals whose mentors had trained very few protégés -- those protégés actually trained more students than expected, a direct contradiction to the rising star hypothesis. But upon closer examination, Amaral and his colleagues noticed a rather strange and pronounced pattern: For mentors that advised at least 10 students in their careers, those protégés that were among the first third of protégés to be advised ended up mentoring significantly more students of their own than expected. Conversely, protégés that were among the last third in a mentor's career ended up advising significantly fewer students. "This was very, very surprising to us," Amaral said. One possible explanation for this effect is what Amaral calls the "mental gap." In essence, researchers late in their careers may find it hard to relate to the struggles encountered by beginning graduate students as they are faced with new responsibilities such as grant reviewing committees and administrative roles at their universities. "As those highly successful mentors move on through their careers, they [could] actually become better and better at mentoring more advanced protégés and less effective at mentoring the ones that are just starting," Amaral said. More research is needed, however, to determine the true cause of the bizarre pattern, he added. In addition, these results may only apply to the field of mathematics. Differences in research and funding in other fields may result in differences in the mentor-protégé relationship. In life sciences, for example, "people that are in that final third of their career oftentimes have more time and perspective and are able to pull themselves away from just doing the research to actually cultivating and developing researchers," O'Neil said. "The other issue for us in the life sciences is what's the impact of [funding]," Lakoski added. The availability of funds, both in terms of new granting mechanisms made available by the National Institutes of Health and other agencies and the relentless oscillations of "fat" and "lean" funding years, is "a big determinant of how many young investigators you can have in your lab," O'Neil said. Additionally, there is the question of "What is mentorship success?" Lakoski asked. While the fecundity metric used in this study "is an interesting approach, it may not cover all the [aspects] of a relationship," she said. Still, Amaral said, this study shows "that you have to think carefully," Amaral said. "If you are a graduate student and you are trying to choose a mentor, it may be very appealing to choose the person who is at the top of their field, who may be later in their career, [but] they might not remember as well what it was like to be a starting scientist." R.D. Malmgren, et al., "The role of mentorship in protégé performance," Nature, 465: 622-6, 2010.
**__Related stories:__***linkurl:Addressing Cultural Caveats;http://www.the-scientist.com/article/display/56180/
[December 2009]*linkurl:Mentoring Magic;http://www.the-scientist.com/article/display/55235/
[December 2008]*linkurl:Management for beginners;http://www.the-scientist.com/article/display/54802/
[July 2008]