Working paper

The Effects of Class Size on the Long-Run Growth in Reading Abilities and Early Adult Outcomes in the Christchurch Health and Development Study (WP 01/14)

Authors: Michael A. Boozer and Tim Maloney

Abstract#

The conflicting evidence over class size effects drawn from observational (or correlational) studies on academic achievement or labor market outcomes has provoked great debate in both academic and public policy arenas. The recent experimental evidence from the United States due to the Project STAR demonstration run in Tennessee in the 1980s has added strong evidence in favor of one type of an effect, whereby initial test score gains were maintained only through sustained smaller class sizes. But this has not resolved the debate by any means. In this paper we utilize the unique feature of the Christchurch Health and Development Study (CHDS) in that children are sampled for extremely long individual histories of their class size experiences as well as their test score and early adult outcomes. We argue that one implication of Project STAR is that only persistent class size reduction policies may have detectable effects, and so the long histories of the type recorded by the CHDS are necessary to detect class size effects. We account for the observational nature of the CHDS (in that children were not randomly assigned to different class sizes) by examining the long-run trends in test score growth, rather than levels.

Consistent with the experimental evidence, we found statistically and economically significant effects of children being assigned to persistently smaller classes on both childhood test score growth as well as on early adult outcomes such as completed education and unemployment experiences. Our analysis points the way towards the unification of experimental and observational evidence on class size effects, as well as highlighting several possible pitfalls in the analysis of observational data on this topic. It also serves as a platform by which further experimental studies on class size might be designed to verify or falsify some of the hypotheses raised in this report, and so further unify the entire literature on this topic