Desempenho escolar no Brasil: efeitos do tamanho da turma, impactos da qualificação docente e a relação com os custos educacionais

Abstract

This thesis is composed by three papers, inherent in Educational Economics, which seek to provide alternative approaches to the analysis of existing relations among some school factors and students‟ educational performance. In the first paper, it was sought to investigate the effects due to the class size, according to the classrooms‟ compositions. Through the matching methodology strategies, the Propensity Score Matching (PSM) and the Generalized Propensity Score (GPS), applied to micro data from SAEB, for the year of 2011,evidence of a negative relation between students‟ grades and the class size was found, especially those from 25 students (or more). The results also suggest that the negative impacts, due to the increase in number of students per class, tend to mitigate in groups formed from the grade homogeneity criteria, and the exacerbation of harmful effects in heterogeneous groups concurrently. In the second paper, aiming to estimate the impact from the quality of the teacher on students‟ Maths school results, the school‟s fixed effect estimators (within estimators) applied to SAEB micro data for the year of 2011 were resorted to. The main findings point that among the variables in the faculty quality vector, the pedagogical practice adopted in the classroom are shown, for instance, to be more relevant than attributes such as academic formation and education, which are traditionally considered as “faculty quality” benchmarks. Furthermore, that the considered traditional teaching strategies, based on repetition and memorization, were found to offer no contribution to students‟ cognitive skills development, regarding Maths teaching. The findings also point that these practices‟ effects are noticeable in higher society stratum, in detriment of groups from less fortunate socioeconomic contexts. In the third paper, a case study repurposed to the public education in Paraiba state is introduced. The main objective of the research was to identify the existence of economies of scale and scope in the productive structure of educational services offered by the state. For this purpose, with a database provided by Paraiba‟s State Secretariat of education, the estimation of a multi-product cost function, transcendental logarithmic type, was adopted as methodological approach. One of the main results suggests that, regarding realized enrolments, the state schools are operating in a segment of economies of scale and that they could reduce costs by offering more than a school level in each school facility. Moreover, the results present evidence that the educational costs yield statistical correlation only with quantitative products such as the number of enrolments, contrary to the products which aim at capturing the offered education‟s quality, such as approval rate and proficiency, for instance.