In this paper we provide a rigorous review of evaluations of education programmes in developing countries. This review is organised to provide guidance to policy-makers by identifying education interventions that are effective, sometimes effective, and not effective and by presenting some information on the cost effectiveness of different types of interventions. We distinguish between interventions' impacts on two types of outcomes: time in schools and learning as measured by test scores. Regarding children's time in school, we find that there are two effective, but admittedly relatively expensive, ways to increase enrolment and attendance: (1) Providing cash transfers conditional on attendance; and (2) Building new schools in communities where schools either do not exist or are far away. We have identified four interventions that have been shown to be effective at increasing test scores: (1) Merit-based student scholarships; (2) Teaching at the right level; (3) Decreasing pupil-teacher ratios; and (4) Building new schools.