The lack of influence of class size on students’ academic performance: empirical evidence for Andalusia

Abstract

Class size has been and continues to be a focus of the Spanish education debate. Most of the literature points towards the negative influence that overcrowded classes may have on students’ academic performance, which has increased the belief that a reduced class size may be better for students’ learning. However, the endogeneity that class size presents has prevented a great part of the research works – which are mostly correlational – to grasp its actual influence. Because of that, we intend to solve this issue by the use of a fuzzy regression discontinuity approach (a combination of regression discontinuity and instrumental variables) using as instrument the class size which schools should have set if they had followed the education legislation. This issue has been analysed for the most populated Spanish region, i.e. Andalusia, using census data for primary and secondary education students, provided by the Andalusian Agency of Education Assessment (AGAEVE) for the academic years 2011/12 and 2012/13. Our results show that class size does not influence students’ administrative scores (in mathematics and reading) in primary and secondary education and that the characteristics of the students in the class may be more important.

Keywords: class size, academic performance, fuzzy regression discontinuity, instrumental variables, primary education, secondary education.