Association for Public Policy Analysis & Management

School-aged children in the U.S. today are more ethnically and racially diverse than they were decades ago. Still, many children receive schooling in racially segregated contexts and the need to foster diversity at schools remains questioned. Testing if there are long-term benefits from having attended diverse schools is therefore a key issue in the educational policy debates. In this paper, we focus on the relationship between attending a sequence of racially diverse schools and students’ choice of STEM as a college major. We focus on STEM because of the pivotal role STEM has on the present and future of economic growth across nations, as well as lasting consequences for individuals’ social mobility and life chances. Drawing from perpetuation theory and theories of cumulative disadvantage, this paper estimates if there is a long-term influence of the sequence of racial composition of schools attended (middle school, high school, 2-year college and 4-year college) on students’ college choice of major (STEM vs. non-STEM). We use the Roots of STEM dataset, which includes information about middle school, high school, two-year and four-year college characteristics and students’ background for the cohort of students who graduated high school in 2004 and entered one of North Carolina's sixteen public four-year universities between 2004 and 2010. We examine outcomes through hierarchical logit modeling. Findings indicate that students who attended a sequence of racially diverse schools are more likely to declare a STEM major than those who did not.