One would expect that family income is an important positive factor in the school attainment of children. However, evidence on this relationship is often tainted by the lack of control for parental ability, since at least a portion of ability is transferred genetically to children. This paper considers empirical strategies that control for both observed and unobserved parental ability. In the end, family income still has a significant effect, which must therefore be causative. It implies that high-ability children in low-income families face binding credit constraints that society may wish to relieve.