Early childhood interventions are often touted as powerful, cost-effective and equity enhancing investments. Critics, however, have been quick to point out that the observed benefits of many large-scale preschool interventions dissipate quickly. Results from the national Head Start Impact Study, a large, randomized trial, showed that at third grade, Head Start participants did not systematically outperform their control group peers on any of the developmental domains examined. Earlier quasi-experimental studies have similarly noted a pattern of “fade out” both among Head Start participants and among preschool participants more broadly. The lack of persisting program impacts is troubling and has meaningful policy implications, particularly as states are expanding their investments in early childhood education. This panel provides new empirical evidence on the persistence of the impacts from large, public preschool initiatives in three unique state contexts (North Carolina, Florida and Tennessee). Using experimental and quasi-experimental methods and rich, student-level data, the studies provide some of the first rigorous evidence about the medium-term (mid/late elementary school) impacts of large state early childhood initiatives. The studies examine whether patterns of fade-out differ across cognitive and non-cognitive outcomes. This is important new evidence as researchers have recently posited that the fade-out of test score gains is not, in and of itself, inconsistent with long-term impacts if the positive effects of preschool operate through benefits to non-cognitive skills. Dodge, Ladd & Muschkin examine the impacts of two major early childhood initiatives in North Carolina on diverse child-level outcomes in fifth grade. Their study is unique in examining the overall impact of these public programs, over and above the impacts for individual participants. Bassok & Miller provide the first empirical evidence about Florida’s Voluntary Pre-Kindergarten program, the largest state preschool program in the country. Leveraging the shock in preschool availability created by the program’s introduction, as well as exogenous county-level variation in preschool slots, they present findings on third grade test scores as well as special education, attendance, and retention in the first three years of school. The final two papers report on results from a large experiment investigating the impacts of Tennessee’s Voluntary Prekindergarten (TN-VPK) on a host of cognitive as well as non-cognitive outcomes including attendance, special education classifications and disciplinary incidents. Lipsey, Hofer & Farran provide the main impact analysis, presenting both the short and medium-term results. Their findings suggest strong initial effects on cognitive outcomes that fade out in the early elementary grades. However, they detect persistent non-cognitive benefits. Springer & Swain leverage the same data, but examine heterogeneity in preschool “effect persistence”, focusing in particular on whether fade-out is less pronounced when children have high-quality elementary school teachers. Their results provide evidence that the learning environment in the years subsequent to preschool plays an important role in maintaining preschool impacts. The session will focus on the policy implications of these four studies. Our two discussants have worked extensively on related questions around preschool fade-out using data from Oklahoma’s Universal Preschool Program and the National Head Start Program.