Abstract
This paper aims to estimate the impact of the frequency of child education in the proficiency of public elementary school students in Brazil, using as proxy the scores of students in Portuguese and Math tests available in Prova Brasil 2011. The Propensity Score Matching methodology was used under the selection hypothesis by observable variables, which aims to find a counterfactual. In order to control characteristics related to the students' performance, linear regressions were estimated by the robust method of least squares weighted by propensity scores. The results point to positive and statistically significant effects on the grades of Portuguese and Math in the 5th and 9th grades for students who began in pre-school and for students who started in kindergarten compared with those who started directly in the primary school.
Keywords:
Early childhood education; Propensity score matching; Least squares weighted by Propensity Score
