Together with colleagues from the QUEST Center for Responsible Research, a new study by CFA postdoc Tamarinde Haven in eLife indicates there to be a greater odds of PhD students sharing data when they had worked under a supervisor that shared data. 

Various codes of conduct and influential commentaries have emphasised the role supervisors should play in socialising PhD candidates into responsible and open research. However, little empirical evidence was available to back this claim. 

Departing from the assumption that if role modelling is important, the study team reasoned that should be possible to discern an association between the supervisor’s engagement in open science practices and the PhD candidate’s engagement in the same practices. 

The analysis, using publication data from 211 pairs of biomedical PhD candidates and their primary supervisors, finds that working under a supervisor that shared data is associated with a greater odds to share data compared to working under a supervisor that did not share data.

There initially seemed to also be an association for publishing open access, but when adding institution to the models, this effect became non-significant. One reason for this could be increased uptake of open access publishing across the Dutch science system. 

Tamarinde is cautiously optimistic about the cross-sectional and non-directional results. Within the open science movement, PhD candidates play an important role in pushing for greater uptake of open science practices — this study may add by showing a new way to examine the role of supervisors in promoting open science. 

You can read the full report here.