THE PURPOSE of this naturalistic experiment was to discover the operative elements of an evidence-based reform effort that accounted for growth in student achievement. The reform was implemented in 13 schools around the United States over the course of two years. The authors studied the impact of both school-level programmatic elements and classroom-level curricular and pedagogical practices. Evidence played a role in two distinct ways: (a) in the research base that staff members were encouraged to consult to shape their local reform efforts and (b) in the data they were provided at key points concerning their progress in shaping the schoolwide reading program, altering classroom teaching practices, and improving student scores on a range of outcome measures. Through HLM analyses of school-level and classroom-level variables, the authors found that success in implementing the reform explained a small but significant proportion of the between school variance in reading growth (as measured by comprehension and fluency scores) when looking across a given year, but it explained a substantial proportion of the between school variance in reading comprehension growth when looking across a two-year period. The teachers in the high-reform-effort schools used more research-based reading instruction practices than teachers in the low-reform-effort schools and made more changes in the direction of research-based practice. Differences between the high-reform-effort and low-reform-effort schools in implementing the various components of the school change framework are discussed in light of the broader research in school reform and national and state policies regarding school change.