This paper presents a longitudinal analysis of cross-national data on armed conflict, state fragility, and enrolment in primary and secondary schooling. The study is motivated by questions raised in the 2012 Human Security Report, which challenges the widely held assumption that conflict is necessarily detrimental to educational outcomes. We use multilevel modelling techniques to determine how conflict and fragility relate to changes in enrolment. Our findings suggest that growth in enrolment is significantly lower in conflict-affected countries but that the effect is dependent upon countries' overall enrolment level. However, when we control for fragility, the effect of conflict is not significant, which is consistent with the Human Security Report's suggestion that fragility is an underlying cause of both conflict and poor educational outcomes. We conclude by discussing the relevance of our findings and challenges for future research on fragility and education.