Three important timing parameters for early childhood education program are intertwined: starting age, program duration and the persistence of program impacts. Our paper uses meta-analytic techniques and cross-study variability in assessment ages to estimate the separate effects of all three time-related components and key interactions among them. Data on 1,978 effect sizes from 117 studies are drawn from a large meta-analytic database being compiled by the National Forum on Early Childhood Policy and Programs. Taken as a whole, effect sizes averaged .27 standard deviations. We find that programs starting before age 3 have effect sizes that average about .10 sd higher than later-starting programs, although this difference was not statistically significant. Effect sizes varied little by program duration. In the case of the persistence of program effects, our point estimates suggest that impacts generally persisted at close to full strength for 1-2 years beyond the end of the programs but at much less than full strength after that. (Author abstract)