Abstract
Contrary to theory and evidence from many other regions, previous studies in sub-Saharan Africa have found surprisingly weak associations between family size and schooling. It is unclear, however, whether these weak results reflect (spurious) limitations in methodology or (real) differences in context. This study uses schooling histories from Cameroon to test four contending methodological and contextual explanations for these weak results: measurement bias, statistical estimation bias, family buffering, and socioeconomic context. We find the strongest support for the last explanation: the relationship between family size and schooling varies with spatial and historical context. This relationship has strengthened within the country over time, and this raises concern about the implications of current demographic transitions on inequality among children.