While much research focuses on how high school size affects student academic performance, little is known about the relationship between size and non-academic outcomes. Students and parents, however, may care quite a bit about non-academic dimensions of pupil wellbeing, such as safe environments, a culture of learning, discipline, student satisfaction and health. In which schools are students happier and healthier? Does school size matter?

In this paper we explore the question of whether small high schools affect a wide range of non-academic outcomes. In particular, we examine the effect of small schools on children’s attitudes along three dimensions: interpersonal relationships, academic expectations and support, and social behavior and safety. In parallel, we consider obesity outcomes. Hence, a major contribution of this paper is the comprehensive set of non-academic outcomes we are able to explore. Moreover, because recent evidence suggests that size, in and of itself, does not create positive learning environments we also consider a relevant dimension of small school’s heterogeneity: school age (Schwartz, Stiefel and Wiswall, 2012). We then separate small schools into new and old small schools to disentangle the unique and distinct effects of school size and age on non-academic outcomes.

We explore these questions using a unique data set of school and student-level data from NYC public high school students entering 9^th grade in 2009/10 obtained from the New York City Department of Education (NYCDOE). These data combine data on demographic and student performance, obesity and fitness, and responses of students on surveys about their schools. We rely on the wealth of student information to control for background characteristics and on the use of instrumental variables to identify the causal school effect on students’ non-academic outcomes.

Briefly, we find that while naïve comparisons indicate students in small schools have better impressions of their school climate but are more likely to be obese, after correcting for selection on unobservables, school size does not matter for any of these outcomes. Contrary to what advocates of school reform advocate, school size in and of itself does not lead to better student non-academic outcomes. These results, however, hide an important source of heterogeneity in school effects driven by school age. In particular, we find that there are marked differences between old small and new small schools as compared to each other and large schools. While boys’ weight is not affected by school size and age, compared to large schools, girls in new small schools are more likely to be obese, overweight and have a higher BMI. Students also report better interpersonal relationships, social behavior and higher academic expectations in old small schools relative to large schools and new small schools.