The treatment effect of attending a high-quality school and the influence of unobservables
This paper studies the effect of attending a high-quality middle school on subsequent educational outcomes. The analysis is based on data from the German socio-economic household panel in which we observe children when they make their secondary school decisions (in the age between 10-12) and later when they self-report on their intentions with regard to their further educational pathway (in the age of 16-17). To identify the treatment effect, we use a regression-control framework as well as an instrumental variable approach (based on local supply of schools). In a second step, we carefully examine the influence of unobservable characteristics in both methods using the new technique proposed by Altonji, Elder and Taber (2005). Our findings suggest that unobservable characteristics are indeed crucial to the validity of the research design. While we find large positive and significant effects of attending a high-quality school, we cannot rule out that the estimates are not in fact driven by selection on unobservables.