Download the Study

Maimonides Rule Redux

Maimonides Rule Redux: Does Class Size Affect Student Achievement?

New Findings, School Reform, June 2017

We use the dis­con­tin­u­ous func­tion of enroll­ment known as Maimonides Rule as an instru­ment for class size in large Israeli sam­ples from 2002-2011. As in the 1991 data ana­lyzed by Angrist and Lavy (1999), Maimonides Rule still has a strong first stage. In con­trast with the ear­lier Israeli esti­mates, how­ever, Maimonides-based instru­men­tal vari­ables esti­mates using more recent data show no effect of class size on achieve­ment. The new data also reveal sub­stan­tial enroll­ment sort­ing near Maimonides cut­offs, with too many schools hav­ing enroll­ment val­ues that just barely pro­duce an extra class. A mod­i­fied rule that uses data on stu­dents’ birth­days to com­pute statu­tory enroll­ment in the absence of enroll­ment manip­u­la­tion also gen­er­ates a pre­cisely esti­mated zero. In older data, the orig­i­nal Maimonides Rule is unre­lated to socioe­co­nomic char­ac­ter­is­tics, while in more recent data, the orig­i­nal rule is unre­lated to socioe­co­nomic char­ac­ter­is­tics con­di­tional on a few con­trols. Enrollment manip­u­la­tion there­fore appears to be innocu­ous: nei­ther the orig­i­nal neg­a­tive effects nor the recent data zeros seem likely to be manip­u­la­tion artifacts.