A new measure of ability peer effects: national exam scores Vs. school GPA, evidence from Portugal

Abstract

We study ability peer effects at the class-level for 9thgrade students in Portuguese state-owned schools. We resort to school-by-year and school-by-year-by-teacher fixed effects to lessen identification threats stemming from unobserved sorting of students to peer groups. We measure students and peer ability with two alternative forms of achievement: teacher assessment along a broader set of subjects and grades and scores in 6thgrade external exams. Our findings suggest positive peer effects when using lagged National Exams and negative peer effects when measured using lagged teacher assessment of students. We discuss our results and provide suggestions for future research.