O EFEITO DA QUALIDADE DA ESCOLA SOBRE O DESEMPENHO ESCOLAR: UMA AVALIAÇÃO DO ENSINO FUNDAMENTAL NO ESTADO DE SÃO PAULO

The research on the relation between school quality and student achievement have many different results, but most analysts conclude that the school quality has little influence on student performance, as compared to background characteristics. However, there are evidences that difficulties using school inputs to represent school quality could underestimate the quality-achievement relation. The alternative methods suggested in this study are independent of school inputs. In both methods we have used data about 4th grade of Elementary School in São Paulo State from SAEB 2001. In the first one, a decomposition of standardized test score inequality - in part explained by school differences and another by background characteristics - indicates that school differences can explain something between 0.0 and 28.4% of Portuguese score inequality and between 8.7 and 34.4% of Mathematics score inequality. The second method estimates the school impact on student achievement by fixed effects, including dummy variables for each school. It's possible to obtain the interpretation from the simulations made that as for Portuguese as for Mathematics the means of SAEB scores can be improved in one standard deviation, which means 3 years of schooling when all students are included. This also happens when only students from public schools are included (three years of schooling for mathematics and 2.4 for Portuguese scores). This study indicates that educational policies and investments can improve student achievement by replicating existing experiences in the present educational system.