Eric Grodsky

In this paper we assess changes in the direct (or secondary) and indirect (or primary) effects of parental education on baccalaureate college attendance and completion for students who complete high school in 1982, 1992 and 2004. We argue that these effects have different implications for our understanding of social stratification and for policies intended to reduce social stratification in educational attainment. Indirect effects, we suggest, operate through the association between social origins and secondary school achievement (as measured by high school GPA, course taking and test scores). Direct effects are then the associations between social origins and postsecondary outcomes net of secondary school achievement. Taking advantage of new methods to decompose the direct and indirect effects of social origins in logistic response models, we find that the baccalaureate college attendance advantage children of more educated parents have enjoyed over children of less educated parents has remained fairly stable over time. However, the degree to which that advantage operates through high school achievement has increased from around 45% to 50% in 1982 to about 60% in 2004. In contrast, the advantage in elite college attendance enjoyed by children of more highly educated parent has increased over time, while the share of that advantage that operates through secondary school achievement has remained stable or even declined. Gradients in baccalaureate completion have been fairly stable according to transcript based measures and trends in the relative contribution of direct and indirect pathways mixed.