Whether supervision of doctoral students is best pursued individually or collectively is a recurring but unresolved question in debates on higher education. The rarity of longitudinal data and the common usage of qualitative methods to analyse a limited number of cases have left the effectiveness of either model largely untested. To assist with overcoming these problems, this paper reports on a study of 145 individuals admitted to a specific doctoral programme between 1991 and 2014. It analyses the effects of either individual or collective supervision during the first year of the programme on the probability of thesis completion and the time to thesis completion. Group means, Cox regressions, Kaplan–Meir curves and Ordinary Least Square regressions are calculated on the basis of the number of months spent by each doctoral student in the programme without defending a thesis. Studied in these ways, it appears that collective supervision in the first year significantly increases the probability of thesis completion and decreases the time to thesis completion. Collective supervision may have this effect as it enhances peer learning, creates a wider academic learning context, allows doctoral students to gradually acquire the values and behaviours of a research practice community and reduces the risk of premature selection of permanent supervisors.