Seminarium: José Montalbán-Castilla, SOFI

Abstract: This paper examines the effects of class size on the learning environment, student achievement, and well-being. Focusing on Madrid, we link administrative data with large scale surveys of students, parents, and teachers, and exploit variation in class size due to a maximum class size rule. We find that smaller classes lead to less classroom disruption, more small-group teaching, and more personalized instruction. However, the effects on student achievement are small and statistically insignificant. We conclude that achievement effects are minimal because smaller classes induce only minor changes in the classroom learning environment, and not because of compensating responses (e.g., by parents). While smaller classes affect the home learning environment (e.g., they result in less parental involvement and less time spent on homework), these effects are small and mediation analyses suggest that they cannot explain the small impact on test scores. Consistent with smaller classes inducing modest changes in both the classroom and home environments, we find that they lead to slight improvements in well-being among students, parents, and teachers.