We evaluate the effectiveness of small high school reform in the country’s largest school district, New York City. Using a rich administrative dataset for multiple cohorts of students, we estimate a model of school outcomes using instrumental variables with multiple types of endogenous school...