Social background's effect on educational attainment: does method matter?

Social background directly impacts educational choice and attainment, but also influences choice and attainment indirectly by affecting school performance. Boudon (1974) described this relationship as primary (indirect) and secondary (direct) effects of social stratification. Based on this approach and Mare¿s sequential transition model, we decompose this impact to analyze these effects¿ relative importance at various stages over the school career. Using Dutch panel data of three school cohorts, we can assess whether primary and secondary effects¿ relative importance has been stable over time. We use different statistical methods to assess the results¿ robustness. Our findings show secondary effects have a decreasing impact at the first transition over time but a rather stable and in some cases increasing impact at the educational career¿s later stages. As a result, the cumulative share of secondary effects on educational attainment is stable over time, at least if one examines the last two cohorts. When using ordinary least squares (OLS) or counterfactual models, secondary effects amount to some 55% of social background¿s total effect. However, using structural equation modeling that allows for taking into account measurement error in performance tests and social background, secondary effects¿ relative importance amounts to some 45%. This result suggests method does matter for numerical closeness. Nevertheless, the findings of all models used in this study point in the same direction and suggest that preferences and expectations of aspiring higher educational levels remain strongly associated with social background.