Early-career researchers are more likely to drop out of academia if they are working with successful mentors who lead large groups. That is according to an international team of researchers who suggest that the effect could be due to mentees having greater competition for a mentor’s time (arXiv: 2208.05304).
Running a large group is often considered a sign of academic success. Indeed, previous studies have shown that academics who are trained in large groups by successful mentors are more likely to be successful and to have more future mentees themselves. However, those studies often looked only at individuals who continued in academia, so it was unclear how “survivor bias” was affecting results.
The latest work quantitatively investigates the advantages and disadvantages of being mentored in large or small groups as an early-career scientist. The authors analysed information on academic genealogy from the Academic Family Tree website and publication data from Microsoft Academic Graph.
Comparing these datasets, they matched the genealogical data of 309,654 scientists with 9,248,726 papers that were published between 1900 and 2021 in physics, chemistry and neuroscience.
After examining the numbers of co-mentees that individuals had, the authors labelled 25% as having been mentored in “big groups” and 25% in “small groups”. They then found that, from the 1950s to the present day, the “survival rate” – or the percentage of those that stayed in academia – was significantly lower for those mentored in large groups compared to small groups. In 1990, for example, the survival rate in physics was 61% for small-group mentees, but only 33% for large-group mentees.
When the researchers considered only individuals who remained in academia, they saw the same effect as in previous research. Large-group mentees, in other words, were more likely to achieve greater academic success, in terms of publications, citations and the number of mentees they went on to supervise.
Making connections
Data scientist and co-author Roberta Sinatra from the Univeristy of Copenhagen suggests that the latest findings, which have yet to be peer-reviewed, could prompt an important discussion.
“The common narrative is that we should increase retention, especially of graduate students, and improve their wellbeing,” Sinatra told Physics World. “Yet the scientific enterprise implicitly promotes high impact, high productivity and publications in top-tier journals. If we truly believe in our stated goals, then we should inspect the reasons for these high dropout rates and promote a more equal distribution of early-career researchers.”
Avoid large groups to be a disruptive scientist
Network theorist Iris Wanzenböck from the University of Utrecht, who was not involved with the latest work, says the results are consistent with her own observations. “These findings confirm that science is a social endeavour, influenced by networks and the quality of connections,” she says. “I think we should be more aware that academics have long-lasting effects on the system by training the next generation. For most, this impact will be much more direct than through their publication or citation numbers.”
