This study uses data from a randomized controlled trial of two treatment conditions (unstructured, free and semi-structured block play) to assess the extent to which executive function mediates intervention impacts on mathematics assessments to approximate a causally informative estimate. Participants for this study included 230 children (50% female) who ranged in age from 36.94 to 63.40 months (M = 52.01, SD = 6.78). All children came from families with low incomes, as defined by Head Start status or eligibility for free or reduced lunch. Research Findings: Our findings demonstrate that executive function mediated the effects of one of the block play conditions (free play with blocks) on children’s improvement of numeracy and number identification, but not the other condition (semi-structured block play). Practice or Policy: Our findings suggest that a mechanism underlying the impacts of free play with blocks on children’s numeracy and number identification is executive function, which contributes to developmental theory around why block play promotes early math. Our findings also support the multifaceted role of incorporating intentional free play with blocks into the daily schedules of early childhood education centers.