Modeling class size effects across the achievement distribution

Previous findings from Project STAR have highlighted the benefits of being in small classes in early grades on average. Here, we examined the differential �value-added� effects of small classes across the achievement distribution. We find that once previous grade achievement and small class membership are controlled for, small class effects are by and large insignificant. Although high-achievers benefited more from small classes in third grade reading, overall, the differential small class effects were not systematic. Findings from longitudinal analyses failed to show that being in small classes for two or more years is as beneficial as being in small classes for only one year. Although the cumulative small class effects are positive, significant, and meaningful in magnitude, they do not indicate which grade (or grades) are the most important