Serving the Poor Differently: The Effects of Private and Public Schools on Children’s Academic Achievement in Basic Education in Mexico

By German Trevino (2015 Doctoral dissertation, Harvard Graduate School of Education):

“Private elementary schools in Mexico are usually seen by wealthy and middle-class families as an alternative to public education. However, private schools have not been seen until very recently as an academic alternative for the poor. In my dissertation, I used data on students from poor families (beneficiaries of Oportunidades program) attending sixth grade of elementary school, who participated in the Quality and Educational Achievement Test assessment 2009 (EXCALE06-2009), to evaluate if there is a private school advantage for the poor in Mathematics in Mexico. I also investigated the extent to what achievement differential is explained by important features of private schools, such as physical resources, school management, teacher quality, teaching practices and classroom organization, and peer group composition. I used propensity score-matching to correct for bias arising from the self-sorting of students into type of schools. I concluded that private schools in Mexico offer a clear advantage for poor students in elementary education, even after accounting for selection bias. On average, private school students who are beneficiaries of Oportunidades program outperformed their public counterparts by 48 test points in Mathematics, or 48% of a standard deviation. The results passed different robustness checks and are technically reliable. The 0.48 sd effect size of private school is larger relative to much of the existing literature, especially if it is compared to the results of true experiments or quasi-experiments of private schools conducted in the U.S. I hypothesize that these large results might have to do in part with two factors: the use of the right counterfactual for this research: poor students attending private urban schools; and the fact that students in the sample attending private schools are all beneficiaries of Oportunidades, a comprehensive poverty alleviation program. This might mean that the treatment under study is more complex than just private schooling. After statistically accounting for selection bias, all of the remaining private school effect is accounted for by identifiable school factors. Peer group composition, school management, teacher practices and classroom organization, are the most important factors explaining the private school advantage in Mathematics in elementary schools in Mexico.”