Date of Award

Summer 8-1-2008

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Geography, Geology, and Anthropology

First Advisor

Brian Ceh

Second Advisor

Robert L. Boyd

Third Advisor

Jay D. Gatrell

Abstract

Because a high school graduate's postsecondary expectations influence his or her chances of attaining a Bachelor's degree within eight years, this indicator deserves more consideration than it receives. Using survey results for the postsecondary educational plans of students graduated from Indiana public high schools between 1996 and 2006, this study identifies a typology of four significantly different patterns of postsecondary plans in Indiana school districts. Research literature in geography, political science, and education recognizes educational outcomes, family demographics, and school districts' financial and socioeconomic characteristics as factors that influence college attendance decisions. Many of them differ significantly by the typology of postsecondary plans. Discriminant analysis confirms that they correctly classify school districts into typology clusters significantly more often than would occur randomly. For analysis, the independent vaiiables comprise three groups: student-related, family-related, and school districtrelated factors. Indiana school districts display distinct distributions of postsecondary plans, which demonstrate significai1tly different relationships to the three groups of factors. This is spatial heterogeneity, or non-stationarity. Regression analysis examines how well each group of factors predicts the mean percentage of graduates planning to attend four-year college. A classic regression model iv of factors identified by the literature serves as a benchmark. A second regression model adds indicator variables for typology classifications. Since the dependent variable and many of the independent variables exhibit spatial dependence, a third model perfom1s spatial regression. All three models yield significant results. However, only the spatial regression model explicitly estimates spatial dependence. The presence of significant spatial variables demonstrates that spatial effects must be included if regression results are to be reliable. Place matters as much as to educational aspirations as do other traditional factors. Finally, spatial regression yields useful school district-level infonnation. Instead of global results, it generates mappable measures of local variation.

Download

Share

COinS