This study was conducted by the Central Region Educational Laboratory (REL Central) administered by Mid-continent Research for Education and Learning to provide educators and policymakers with rigorous evidence about the potential of Classroom Assessment for Student Learning (CASL) to improve student achievement. CASL is a widely used professional development program in classroom and formative assessment published by the Assessment Training Institute of Pearson Education. Schools were recruited from across Colorado to participate in the study. Colorado was chosen as the target state primarily because it has one of the largest populations in the Central Region from which to recruit schools and because its statewide achievement test is vertically scaled. This cluster randomized trial of the CASL professional development program had sufficient statistical power to detect an impact of at least 0.25 standard deviation on student achievement. An intent-to-treat analysis was conducted to estimate the impact of CASL on student achievement; all schools were included in the analysis and were analyzed as randomized regardless of the level of implementation fidelity. Analysis did not reveal a statistically significant impact of CASL on the school-level average mathematics achievement of grade 4 and grade 5 students. Results from sensitivity analyses revealed that the impact estimates on student achievement were robust to decisions regarding the inclusion of covariates, estimation method, and the treatment of missing data. In other words, design and analysis decisions made by the research team did not change whether the impact results would have been statistically significant. Appendices include: (1) Power analysis; (2) Response rates by data collection wave, instrument, and experimental group; (3) Data collection instruments; (4) Development, reliability, and validity of teacher outcomes; (5) Teacher Assessment Work Sample; (6) Impact analysis models; (7) Calculation of effect sizes; (8) Treatment of missing data; (9) Variance components estimates and intraclass correlations; (10) Raw means and standard deviations; and (11) Complete mixed model results. (Contains 1 box, 4 figures and 56 tables.