What makes a productive Ph.D. student?

This paper investigates the impact of the social environment to which a Ph.D. student is exposed on her scientific productivity during the training period. Vertical and horizontal relationships depict the social environment. Vertical relationships are those supervisor-student, while horizontal relationships are those student-peers. We characterize these relationships by assessing how the supervisor's and peers' biographic and academic characteristics relate to the student's productivity as measured by the publication quantity, quality, and scientific network size. Unique to our study, we cover the entire student population of a European country for all the STEM fields. Specifically, we analyse the productivity of 77,143 students who graduated in France between 2000 and 2014. We find that having a female supervisor is associated with a higher student's productivity as well as being supervised by a mid-career scientist and having a supervisor with a high academic reputation. The supervisor's fundraising ability benefits only one specific dimension of the student's productivity, i.e., the student's work quality. Interestingly, the supervisor's mentorship experience negatively associates with student's productivity. Having many peers negatively associates with the student's productivity, especially if peers are senior students. Having female peers positively correlates with the student's productivity, while peers' academic status shows mixed effects according to the productivity dimension considered. We find results heterogeneity when breaking down our sample by field of research.