A skeptic's guide to Project STAR

Class size reduction (CSR) is gaining popularity in the US Supporters of CSR often rely on positive results from Project STAR to advance the policy Using the data from STAR, this study identifies two types of schools, depending on the academic performance of students of small and regular classes, and then it examines the distribution of student and teacher characteristics of small and regular classes in both types of schools The main finding is as follows The distribution is uneven in such a way as to raise the achievement of students of small classes in one type of schools and to lower the achievement of students of small classes in the other type of schools The finding casts serious doubt on the claim of randomness in STAR and potentially its results