Private, Private Government-Dependent and Public schools. An International Efficiency Analysis using Propensity Score Matching

This paper aims at estimating the effect on achievement of various types of schools: private, private but government-dependent and public ones. It is based on the analysis of Math, Science and Reading test scores of 15 year-olds students surveyed in 2002 across OECD and non-OECD countries. The estimation of the effect of private vs. public school attendance may be biased by the existence of confounding factors. An obvious start is to use standard (OLS) models to isolate the effect of private/public status from the other determinants of achievement like family resources or socio-economic background. But OLS estimates are highly dependent on the validity of the linearity assumption i.e. that the effect of school type is uniform across the distribution of covariates. Hence, the rational for using non-parametric propensity score matching. The main result is that in most countries examined, the type of school has so statistically significant impact on achievement. There is a small group of countries where students attending private schools (UK, Brazil) or private government-dependent schools (French- Speaking Belgium, France and Ireland) clearly perform better than those attending public schools. But there are also cases like Switzerland and Austria where private schools appear less efficient than public schools.