Circles of Influence

de Chad Joseph D'Entremont

A major focus of charter school research has been the potential impact of increased school choice on student sorting by race/ethnicity and socio-economic status. Researchers have argued that charter schools may increase segregation by allowing families to separate into more homogeneous school communities. Yet surprisingly, little attention has been paid to the role charter schools themselves may play in determining student enrollments. Emerging evidence suggests that charter schools may frequently take into account nearby educational and demographic characteristics when choosing school locations and avoid neighborhoods with high-proportions of at-risk students who are more costly and challenging to educate. While this behavior is not incompatible with traditional notions of how education marketplaces function, such "positioning strategies" serve as important reminders that charter schools benefit most by locating in areas where they hold clear competitive advantages, not necessarily areas of greatest demand or educational need. In this dissertation, I examined the potential for New Jersey charter schools to effectively distribute educational opportunities to all students, particularly those most frequently targeted by previous approaches to school reform, across varied and often segregated landscapes. Drawing on rational choice theory and previous research into the profit maximizing behavior of firms, I argued that charter schools have strong incentives to locate in areas that allow them to effectively balance consumer demand with the potential negative effects of increased competition in high needs areas. I used geographic information systems (GIS) and logistic regression to map the location of charter schools in New Jersey and examine potential associations with supply side factors. New Jersey's charter school supply showed two distinct clustering patterns. First, charter schools tended to circle Abbott districts or low-performing school districts in a narrow five mile band characterized by greater educational need and, presumably, consumer demand. Second, charter schools in Abbott districts tended to circle, but not locate within, neighborhoods with higher levels of educational and economic disadvantage, and particularly neighborhoods with higher proportions of African American residents. Logistic regression confirmed statistically significant associations among charter school clusters and proxies for market demand, educational need, and neighborhood diversity, although estimates for race/ethnicity were less conclusive. Further analysis indicated that observed clustering patterns were primarily driven by more market savvy or entrepreneurial schools partnered with charter management organizations (CMOs).