The effects of individual characteristics on the distribution of college performance

Abstract

Interactions between observed and unobserved determinants of educational success imply that the former have a heterogeneous effect on performance. To quantify these interactions, a quantile regression model is estimated using a database of students at public universities in Argentina. The empirical results show that all factors which contribute positively to performance are stronger in the lower end of the distribution. Hence, policies that enhance the possibilities of students in this part of the conditional distribution have the dual effect of increasing absolute performance and reducing disparities due to their stronger effect in this group of students.