![Sample our Education journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5932/TOC-Subject-Sample-2021-Education.jpg"})
Abstract
This study uses a comprehensive dataset on secondary school students in Brazil to examine the impact of private school enrollment on educational attainment in São Paulo. The results show that private school students (across all levels of tuition) perform better than their public school counterparts on Brazil’s high school exit exam, even after reducing sample selection bias via a variety of propensity score matching models. Ultimately, the author concludes that the significant achievement gains provided by low-fee and high-fee private secondary schools in Brazil strengthen the demand for private schooling despite recent increases in public spending.
KEYWORDS:
Notes
1. There is no consensus on an exact definition for “low-fee” across the literature. In prior work, Heyneman and Stern (Citation2014) have suggested that low-fee schools should be those with tuition fees less than one half of the monthly minimum wage for a given country. This cutoff was tested for the present study but the final tuition level for reported analyses was higher due to sample size restrictions (as explained in the Data section of this article).
2. Convention on the Rights of the Child. For an explanation of the CRC’s measures of educational quality, see CRC Article 29.
3. This has, not surprisingly, made it difficult for researchers to undertake studies that assess the quality of the private school sector in Brazil.
4. Estimates were also run for schools with tuition of no more than half the monthly minimum wage. The coefficients on the main variable of interest were similar in magnitude but imprecise due to the small sample of schools.
5. Differences in writing exam scores are smaller than those for objective scores but the overall trends remain the same across both outcomes.
6. Writing scores were chosen because the overall effect was smaller and therefore more susceptible to attenuation based on unobserved confounders.