State-funded prekindergarten (preK) programs are increasingly common across the country. This study estimated the effects of eight state-funded preK programs (Arkansas, California, Michigan, New Jersey, New Mexico, Oklahoma, South Carolina, and West Virginia) on children's learning using a regression discontinuity design. These programs vary with respect to the population served, program design, and context. Weighted average effect sizes from instrumental variables analyses across these states are 0.24 for language (vocabulary), 0.44 for math, and 1.10 for emergent literacy. Differences in effect sizes by domain suggest that preK programs should attend more to enhancing learning beyond simple literacy skills. State preK programs appear to differ in their effects. We offer recommendations for more rigorous, regular evaluation. (author abstract)