Understanding the private-public school performance gap in PISA

Abstract

We analyse the PISA-reported convergence in the performance of private and public schools in Portugal. When PISA sampling weights are used, the number of students enrolled in those types of schools and specific grades/tracks of study differs significantly from official population figures. To account for those differences, we apply a post-stratification adjustment; however, sample sizes are small, resulting in estimates with low precision for several subgroups. We propose recommendations for improving the handling of these issues in future PISA cycles. In an additional analysis, we also account for changes in the composition of the student population. When all factors are considered, the convergence in scores is far less impressive than reported. For instance, in Science, after adjusting the sampling weights and removing population composition effects, the reported convergence of 46 points between private and public schools from 2015 to 2018 amounts to only 9 points. The decomposition and sample adjustment methods used in this paper can be easily adapted to other contexts.