This study proposes and tests a multilevel structural model of school context, composition, and school leadership on school instructional practices and outcomes in elementary schools in a western state in the United States. We focus on direct and indirect relationships implied in our proposed model using an ''added year of schooling'' in reading and math as our primary school-level outcomes. Added-year effects, which result from a regression discontinuity design, represent a relatively new approach for describing how school factors influence outcomes. Our results suggest that, net of context and composition factors, improvement-focused school leadership directly affected subsequent school instructional practices and, in turn, instructional practices affected addedyear outcomes. We discuss the findings in terms of their theoretical and practical implications for conducting further educational effectiveness research.