Public prekindergarten (pre-K) programs have been a recent focus of policy and research attention, in part because of their empirically documented, positive short-term impacts on child cognitive development and school readiness. However, no studies have explored factors that might explain variation across schools in public pre-K impacts. The current study examines the Tulsa Public School pre-K program's impacts on children's letter-word identification, spelling, and applied problem-solving skills at kindergarten entry. Findings reveal substantial across-school variation in treatment impacts, associated with variation in instructional support. Results are robust to sensitivity checks and are discussed alongside implications for policy and future research. (author abstract)