Bernardo Pincheira

Peer effects in the classroom: Evidence from a natural experiment

Abstract
The causal effect of classmates on students’ academic performance, known as peer effects, is a question still open to debate in the field of education economics. In the paper I use the 8.8 earthquake that struck Chile in 2010 as a natural experiment for exogenous variation in peer composition, using the fact that the earthquake hit a random area of the country and forced some students to move into new schools for non-academic reasons. I use OLS and instrumental variables econometric specifications, with data of students observed in 2010, in affected or non-affected areas, to answer this question. My results show that the peer effects are positive and significant for students in both fourth and tenth grade. An increase of one standard deviation in the average score of the peers has an effect between 0.14 and 0.22 of a standard deviations in the student's own score. In addition, using quantile regressions, I find evidence of nonlinearities in the effect that potentially allow for within school sorting that might reduce inequality of performance between students and marginally increase average performance according to Monte Carlo simulations.

Lunch from 12.30pm