A Spatial Theory of School Closure: An Examination of School Closure in America

Abstract

This dissertation seeks to uncover the primary factors associated with public school closure in the United States during the five year period 2007 – 2012. Specifically, it addresses three questions related to school closures. First, what are the most important socio-economic factors associated with an individual school being closed? Second, are school districts influenced by their neighbors in making school closure decisions? And finally, what is the role of the spatial organization of cities in shaping school closure outcomes? The econometric analysis is conducted using a multilevel model to estimate the correlation of several variables hypothesized to have a relationship to the school closure process. The random effects model estimates the probability of each school district closing a school. These probabilities are mapped and analyzed using the Getis-Ord Gi* procedure for exploring data that are clustered across geographies. Overall, school enrollment, race, and poverty were the most consistently correlated to school closure. School districts also appear to influence one another, but this is a local process of spatial influence. The final question is addressed by reducing the data set to urban districts only. Relying on an urban classification system developed in the late 1980s, the model attempts to account for the impact of the spatial arrangement of urban space on the school closure process. The results find a substantial effect of spatial form of so-called “industrial” cities on the school closure process. Finally, this project examined differences in the spatial distribution of closed schools compared to new schools. School closure disproportionately impacts low-income and minority students of color; it is primarily an urban process; and schools in closer proximity to the metropolitan principle cities are increasingly likely to close. New schools, in contrast, are located in urban regions outside the metropolitan principle cities, but new schools are highly concentrated, primarily in areas of higher relative incomes and lower shares of African American students. Overall, this project establishes the geography of school closure in the era of No Child Left Behind. It concludes that school closure is indeed a process that is grounded and shaped by racial and economic institutions.