Classroom homogeneity and the distribution of student math performance: A country-level fixed-effects analysis

Is the homogenous grouping of students by ability conducive to learning for all students alike, or does it affect students of different ability levels differently? To address this question, I compare the distributions of math performance for students between Grade 8 and Grade 4 across countries with different levels of between-classroom ability grouping, controlling for country-level unobserved heterogeneity using a fixed-effects model. Homogeneous grouping, relative to heterogeneous grouping, is found to have no significant impact on mean performance, but it does increase performance inequality by benefiting the high achievers at the expense of the low achievers.