In the past 15 years, four-year-olds' enrollment in state-funded pre-kindergarten in the United States has doubled, and advocates have pushed for further expansion. Although research has shown that pre-K programs can have important benefits, most existing studies have focused on small or state-specific programs that may not generalize to other areas or contexts. The uniqueness of our paper is its scope: our data cover the last two decades, span nearly all states, and allow for intrastate variation. For the average state program, we find no evidence of effects on the average student's test scores, assignment to special education, or grade retention. Our estimates rule out pre-K impacts as small as 2 percentiles. However, these averages conceal some important heterogeneity. In states previously found to have high-quality pre-K, we find positive effects on math test scores. For majority-black districts, the average pre-K program has large effects on math and reading. (author abstract)