Ethiopia has made tremendous progress in expanding primary education since the mid-1990s. Yet the country continues to face daunting challenges, as it works to meet the goal of primary education for all by 2015. Given the limited resources available for meeting this challenge, it is imperative that resources be spent knowledgeably and wisely. This paper aims to inform policy design by examining the lessons to be learned from three recent Ethiopian household surveys regarding the key barriers to primary schooling, and discussing the implications of the results for policy. All three datasets employ large, nationally representative samples, but bring differing strengths to education policy analysis. The Welfare Monitoring Survey/Household Income and Consumption Expenditure Survey of 1999/2000 contains a rich set of variables describing households’ distance from various types of economic and social infrastructure, allowing assessment of the role that geographic barriers play in primary school enrollment. It is also the only dataset to contain a good measure of household consumption expenditure. Thus the roles that low income and geographic distance from school play in preventing children from attending primary school may be examined simultaneously. The Labor Force Survey of 1999 allows of children’s involvement in work as well as school, shedding some light on opportunity costs of children’s time. The Demographic and Health Survey of 2000 allows more detailed examination of the potential importance of language and parental attitudes in explaining differences in enrollment rates. The research generates both policy-relevant insights, and suggestions for modifications to future Ethiopian data collection efforts that would enhance the ability to draw inferences of relevance to education.