Creating the right environment for PhD productivity

This is a guest post by Alberto Corsini (Université Côte d’Azur, France), Michele Pezzoni (Université Côte d’Azur, France), and Fabiana Visentin (Maastricht University, the Netherlands), based on their recent paper on PhD productivity, in the journal Research Policy.

Today’s science relies increasingly on PhD students’ work. PhD students, through their publication activity, play a fundamental role in contributing to the advancement of the scientific knowledge frontier. Despite their significant contribution to science, few studies have investigated the work environment characteristics that make a PhD student productive. Extant studies lack an encompassing view of the work environment having focused on specific characteristics, such as the supervisor’s biographic profile, research team composition, and affiliation quality. Moreover, those studies have focused on specific disciplines or on relatively small samples of PhD students affiliated with one or a handful of highly reputed universities. An overview of the broad set of work environment characteristics for a large set of individuals is missing, making it challenging to understand just what does make a PhD student productive.

In our recent Research Policy article – “What makes a productive PhD student?”, we explored this topic. We looked at the PhDs’ productivity in terms of publication quantity and quality, and scientific network size, and identified the work environment characteristics that benefit the students’ productivity during the PhD training period. We considered a broad set of work environment characteristics, including a detailed set of supervisor and peer characteristics.

We analyzed the entire population of French PhD students who graduated between 2000 and 2014 in STEM disciplines. On a population sample of 77,143 students, we found that having a productive, mid-career, low-experienced female supervisor who benefits from a national grant is positively associated with PhD student productivity.

Furthermore, having productive freshman peers and at least one female peer is positively associated with productivity.

In addition to intuitive results, such as a positive correlation between having a productive supervisor and peers and student’s productivity, we identified novel results. In particular, we want to discuss here three findings that offer helpful insights to supervisors in boosting their students’ productivity.

First, PhD student productivity is negatively associated with the number of students being simultaneously mentored by a supervisor. This result can be interpreted as less effective supervisor mentorship activity in large teams, due to insufficient time and attention allocated to each student. Interestingly, this result holds the same across disciplines.

Second, the supervisor’s experience, measured by the number of students they have mentored in the past, is negatively associated with the student’s productivity. This result contradicts our initial expectations that more experience might have led the supervisor to be a better mentor. Although unexpected, this result can be interpreted as the supervisor’s tendency to be more supportive when newer to supervision, and feeling more responsibility toward the student as a new supervisor. Also, this result is similar across disciplines.

Third, gender in the work environment matters. Having a female supervisor is positively associated with PhD student productivity, especially in engineering. Interestingly, engineering is the discipline with the lowest presence of female supervisors. Despite our empirical evidence documenting a gender-related effect on student productivity, we still do not know much about how female supervision is beneficial. Digging into these mechanisms requires us to complement our work with interviews with students and supervisors to identify differences in mentorship styles.

Our findings provide hints to supervisors and institutions who want to supervise their students more effectively. According to our results, supervisors should consider limiting the number of PhD students supervised simultaneously, large student groups are detrimental to student productivity. Moreover, inexperienced supervisors should not refrain from taking students. Indeed, the high level of attention devoted to the first cohorts of PhD students will benefit their productivity. Finally, attention should be addressed to identifying the best practices adopted by female supervisors to understand the mechanisms that enhance students’ productivity. Our results offer a lesson also to PhD students who can leverage environmental factors to increase their productivity.