Competências socioemocionais e desempenho educacional no Brasil : uma análise de equações estruturais e pareamento com os dados do PISA 2012

Abstract

This study used the 2012 Programme for International Student Assessment (PISA) data sets to estimate the effects of socio-emotional skills on mathematics performance of Brazilian students through a combination of two techniques: structural equations modeling and propensity scores matching. It also investigated covariates related to student individual characteristics, socioeconomic status, school characteristics, and parental attitudes towards education. The conceptual framework of determinants of school performance provided by Francisco Soares (2008), the theory of planned behavior of Icek Ajzen (1991), and the “Big Five” personality traits dimensions developed by Lewis Goldberg and others guided the construction of the structural equations models as well as the analysis with propensity scores matching. The social-emotional skills construct consisted of five observed measures: perseverance, openness for problem solving, attitude towards school, mathematics self-efficacy, and mathematics work ethic. The sample consisted of 5,074 students from 779 public and private schools. The results achieved with the full structural equation model indicated positive and significant effect of socio-emotional skills on mathematics performance, second only to the average socioeconomic status of the school and the student's family. Propensity scores matching analysis suggested moderate effects size of those skills on mathematics scores according to the type of school where the student was enrolled (public or private). These results reinforce the moderating role of socio-emotional skills in school performance.