In this study, a school district adopted Kindergarten Peer-Assisted Learning Strategies (K-PALS), a scientifically based, class-wide peer-tutoring program for reading. Sixteen new K-PALS teachers were assigned randomly to receive ongoing support from a university expert or from experienced K-PALS teachers within the district. K-PALS teachers who received university support implemented K-PALS with somewhat higher (d =.41), but not reliably different, fidelity compared to those who received district support. K-PALS teachers' student reading outcomes did not differ significantly based on whether they received university or district support. However, K-PALS teachers' students reliably outperformed historical controls on beginning reading measures (d =.24 to 1.29). Implications for further research and for schools' adoption of K-PALS are discussed.