Implementing educational innovations through facilitators: Cascade model works, yet school contexts matter

The cascade model, in which facilitators provide professional development to teachers who then implement innovations, can result in diluted impact along the cascades. Effectiveness has rarely been shown for the level of student outcomes and supportive conditions, such as teacher collaboration and principal leadership, are needed. This study examines whether implementing a research-based, language-responsive mathematics innovation through a cascade model impacts student outcomes and whether teacher collaboration and principal leadership support implementation. Student outcomes were assessed before and after the teaching unit in view in an intervention group and a control group ( n = 479), with school context conditions reported by teachers. Multilevel models revealed that implementation in the cascade model was effective in improving student outcomes. Principal leadership positively predicted outcomes, especially in the control group providing mathematics instruction as usual, while teacher collaboration was a negative predictor. Implications for future research and the design of professional development programs are discussed. • Investigates implementation of innovation in cascade model. • Students in intervention group show higher outcomes than in control group. • Principal leadership positively predicted outcomes, less in intervention group. • Surprisingly, teacher collaboration was a negative predictor of student outcomes.