What kind of scholar makes the ideal PhD supervisor? The answer, according to a major study, is a mid-career female academic with a national research grant and little supervision experience.

In a paper in Research Policy that explores the characteristics of successful PhD supervisors, academics analysed the publication records of more than 77,000 science PhD graduates to see if these traits helped their former students to succeed, with journal outputs seen as a strong indicator of career success.

The study, which considered the research publications of every science, technology, engineering and mathematics PhD graduate in France between 2000 and 2014, concludes that ¡°having a productive, mid-career, low-experienced, female supervisor who benefits from a national grant is positively associated with the student¡¯s productivity¡±.

¡°On the contrary,¡± it continues, ¡°having a supervisor with long mentoring experience and a supervisor in early- or late-career phases is associated with lower student productivity.¡±

ÖйúAƬ

Having a highly published supervisor was also a key trait for students who had also published widely, while a supervisor¡¯s grant success with national bodies also mattered, says the study, written by academics from the University of C?te d¡¯Azur, Sciences Po, Bocconi University and Maastricht University.

Surprisingly, however, holding a European Union grant ¡°negatively relates to the student¡¯s publication quantity and network size¡±, which ¡°might be explained by the additional time spent by the supervisor managing the EU grant¡±.

ÖйúAƬ

Against their expectations, the researchers also found that supervision experience did not correlate with high research productivity among PhD graduates, with a ¡°student mentored by an experienced supervisor [having] fewer papers published, citations received, and a smaller collaboration network¡±.

¡°We interpret our finding as the supervisors¡¯ tendency to be more supportive to students when they are at their first experience as thesis directors,¡± explains the study.

With this in mind, universities and funders should think about expanding supervision opportunities to ¡°scientists who have never mentored PhD students, [who] should be incentivised to start the mentorship activity¡± if they wish to produce more productive PhD students, the study says.

However, academics should be required to have a solid number of publications before they were allowed to take on graduate students given the link between their own research quality and their students¡¯ work.

ÖйúAƬ

In France, where scholars must gain a license, known as a habilitation, mainly based on the scientific output, before mentoring PhD students, ¡°raising the threshold for obtaining the habilitation would ensure supervisors with a higher number of publications and, according to our results, more productive students¡±, says the paper.

jack.grove@timeshighereducation.com