Using a cluster randomization design, schools were randomly assigned to implement Success for All, a comprehensive reading reform model, or control methods. This article reports final literacy outcomes for a 3-year longitudinal sample of children who participated in the treatment or control condition from kindergarten through second grade and a combined longitudinal and in-mover student sample, both of which were nested within 35 schools. Hierarchical linear model analyses of all three outcomes for both samples revealed statistically significant school-level effects of treatment assignment as large as one third of a standard deviation. The results correspond with the Success for All program theory, which emphasizes both comprehensive school-level reform and targeted student-level achievement effects through a multi-year sequencing of literacy instruction. (Contains 5 tables.)