<個人論文>準実験的手法による学級規模効果研究のメタアナリシス

概要

論文の詳細を見る
From 1990's the empirical methods by quasi-experimental designs have grown, many papers on class size effects have been published. However these estimated magnitudes of the class size effects are varying and mixed. This paper aims to review the literatures systematically by the meta-analysis. First, we synthesize the previous researches by quasi-experimental design such as instrumental variable regression, regression discontinuity design and difference-in-differences. The estimated mean class size effect is +0.01 SD per student reduction. Second, we analyze the factors of heterogeneities of the class size effect. The nonlinearity of the class size effects and the factor about the policy on teacher salary are examined.