ECON 151 Lecture Notes - Lecture 44: Thomas Piketty, Kindergarten, Randomized Controlled Trial
Document Summary
Smaller class size associated with better performance. Smaller effect than star: results in star goes from . 32 sd to . 66 sd while israel show around 30 sd. Like in star larger effect for pupils originating from poor background results might depend on initial class size. Similar results were found in developed countries with angrist lavy methodology : Piketty & valdenaire (mimeo 2008) - france. Fredriksson et al. (qje 2013) - sweden; Chetty et al. (qje 2008) - us. Urquiola (2006) finds large effects in bolivia. Class size is found to increase performance. Effect is between -1. 5 and -2. 5 sd per additional student. + 10 students=> . 15 sd - . 25 sd. Yet, costly policy: 1000 euro/per student=> 2-7. 5 pp sd. Costly small class size has marginal positive impact. Since primary school education cost 7000, 1 15% increase in the overall cost of education would only marginally increase performance. Class size should be concentrated on very deprived schools.