Since World War II, efforts to improve schools have numbered in the thousands. Most efforts have concentrated on improving the curriculum materials used in schools or on "training" teachers in new instructional methods. Many of these efforts have gone under the banner of "building instructional capacity," a term that for decades has been featured prominently in conversations about educational reform. Unfortunately, three decades of research has found that only a few interventions have had detectable effects on instruction and that, when such effects are detected, they rarely are sustained over time. A review of research and professional experience with school improvement suggests several explanations for these disheartening findings. One is that schools are complex social organizations situated within, and vitally affected by, other complex social systems including families, communities, and professional and regulatory agencies. The larger social environment of schools constrains and shapes the actions of teachers, students, and administrators, often in ways that greatly complicate the work of school improvement. Challenges to school improvement are particularly acute in highpoverty settings where recruiting wellqualified teachers is difficult and where the emotional and health problems of students often deflects attention to educational issues or impedes work on them. As a result, many researchers now believe that school improvement involves much more than efforts to change interactions occurring within schools. To succeed, school improvement interventions also must attend to the complex relationships that exist among intervention agents, schools, and their social environments.