The purpose of this study was to make use of indicators of level of implementation collected during the enactment of Oregon Reading First in order to examine whether variation of implementation of the components of the Schoolwide Reading Program predicted better outcomes for students and schools. In particular, the aim of this study was to determine the extent to which each of three different types of measures of implementation fidelity, as well as a combined index of these measures, explained school-level variance in student improvement in 34 schools participating in the Reading First program. Hierarchical linear modeling was utilized to predict reading performance and growth on oral reading fluency and overall measures of reading performance. Mixed results, at best, were found when analyzing this association. In both second and third grades, one of three implementation indices and a composite total of all three measures were statistically significant but small predictors of oral reading fluency growth. However, this relationship was offset with the removal of one outlier school. Implementation threshold effects are discussed as a possible cause of nullification. No statistically significant relationships were found between implementation fidelity measures and overall reading outcomes directed at reading comprehension. Although not a focus of the study, school-level demographic characteristics including special education status and limited English proficiency appeared to explain significant differences between schools despite the use of evidence-based practices and strong support for implementation of these practices.