Working Paper: CEPR ID: DP18564
Authors: Michela Tincani; Enrico Miglino; Fabian Kosse
Abstract: Exploiting a randomized control trial and a dynamic structural model, we provide evidence on the impact of preferential college admissions in Chile on education outcomes. The college admission policy (PACE) targeted disadvantaged students scoring 1.5 standard deviations below regular college entrants on high school tests. We constructed a 9-year-long longitudinal dataset on 9,006 students linking detailed administrative records to survey data. We show that PACE increased first-year college enrollment by 3.1 percentage points and the effect shrank to 1.1 in the fifth year. The policy decreased the pre-college effort of students, likely due to belief biases about their absolute and relative ability. Using simulations from a dynamic structural model, we show that eliminating the pre-college belief biases would improve the college preparedness of college entrants. Our results demonstrate that expanding admission advantages to very disadvantaged populations can improve their college attainment, but college preparedness matters and it responds to incentives shortly before college.
Keywords: College Admissions; Education Outcomes; Disadvantaged Students; Randomized Control Trial
JEL Codes: I23
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| belief biases (G41) | precollege effort and achievement (A21) |
| study effort and achievement in high school (I21) | standardized tests performance (D29) |
| study effort and achievement in high school (I21) | GPA in core subjects (A21) |
| eliminating belief biases (D91) | college preparedness (A21) |
| PACE policy (E64) | first-year college enrollment (I23) |
| first-year college enrollment (I23) | continuous enrollment or graduation after five years (Y40) |
| PACE policy (E64) | study effort and achievement in high school (I21) |