The effects of class size on cognitive and non-cognitive abilities(in Japanese)

The trend towards class size reduction has accelerated over the last decade. However, it is not clear whether small class sizes actually increase the quality of education enough to justify the downsizing cost. This paper intends to capture the effects of class size by using an education production function and by estimating the effects focusing on the following two points. First, we evaluate various aspects of class size effects. This paper uses TIMSS2003 Mathematics and Science data for Japanese public schools’ 8th grade students to estimate the effects of class size on academic achievements and four non-cognitive ability measures (“High motivation”, “Confidence”, “Utility”, “Belong”), which are estimated by using factor analysis to allow quantitative interpretation. Second, we control for endogeneity of class size that results from non-random assignments of students to class. This paper uses the Regression-Discontinuity-Design method which eliminates such biases. We obtained the following two results. First, we did not find a significant relationship between reductions in class size and academic achievements in both subjects. Second, regarding the non-cognitive ability measures, we found that in mathematics “Confidence” increases with class size reduction.