Do the short-term effects of a high-quality preschool program persist?

•For two cohorts, we compare third-grade test score outcomes of children who previously attended pre-K in Tulsa Public Schools with outcomes of similar children who did not attend.•We use boosted regression and propensity score methods.•For the early cohort, we find no evidence of persistence of early gains.•For the late cohort, early gains persist in math, for boys but not for girls. Early gains do not persist for reading.•We discuss possible reasons for the pattern of findings, but our study design does not allow us to identify the causal mechanisms of persistence.

We investigate the persistence of short-term effects of a high-quality school-based pre-kindergarten program in Tulsa, Oklahoma. We analyze third-grade reading and math scores for two cohorts of students eligible to participate in pre-kindergarten in 2000–2001 and 2005–2006, using boosted regression and propensity score matching to select a comparison group of local students who did not participate in the pre-K program. For the early cohort, we find no evidence of persistence of early gains. For the late cohort, we find that early gains persist through third grade in math but not reading, and for boys but not for girls. We discuss possible reasons for the pattern of findings, though our study design does not allow us to identify the causal mechanisms of persistence.