This paper aims to summarize the unexplained propensity of children to engage in work, school, or neither. After controlling for a wide range of determinants of child labor, schooling, and idleness, we estimate a hierarchical model that allows for heteroskedastic, spatially correlated random effects. We use the posterior distribution of ranks of random effects to capture social norms toward children’s activities in each district and thus identify those Indian districts where social attitudes favor education and oppose child labor and idleness. We propose that government intervention be targeted at districts with pro-schooling, anti-child-labor, and anti-idleness social attitudes if limited government resources necessitate implementing minimal cost policies that have the greatest potential to succeed.