CAN ANALYZING SPATIAL RELATIONSHIPS THROUGH GEOGRAPHICALLY WEIGHTED REGRESSION IMPROVE OUR UNDERSTANDING OF LOW SCHOOL ATTAINMENT? A GIS-BASED ANALYSIS OF CENSUS AND ACS DATA

Authors

William England, University of Nebraska-LincolnFollow

Date of this Version

10-2014

Document Type

Article

Comments

A DISSERTATION Presented to the Faculty of The Graduate College at the University of Nebraska In Partial Fulfillment of Requirements For the Degree of Doctor of Philosophy, Major: Educational Studies, Under the Supervision of Professor Edmund T. Hamann. Lincoln, Nebraska: October, 2014

Copyright (c) 2014 William R. England, Ph.D.

Abstract

In this study I present a relatively new technique for analyzing a recurring problem in our communities. Using a set of innovative and relatively new modeling methods, I demonstrate ways in which it is possible to directly account for, capture, and visualize the spatial variability in the relationships between U.S. Census data from 1990 and the recent low-school-attainment landscape in both the Omaha and Lincoln Public School (OPS) districts in Omaha and Lincoln, NE. Low school attainment in adults is a correlate of a host of troubling health and economic factors, which, in turn, have an impact on a child's school performance and eventual school attainment. Disrupting this trend is (and has been) the focus of much research because not only is low school attainment predictive of a host of concerning variables, school attainment also has a tendency to persist from generation to generation. In addition, areas of an urban environment characterized by low school attainment seem to remain geographically stable over long periods of time. However, traditionally, researchers modeling the relationships associated with school attainment draw conclusions based on techniques that rely on global inferences (e.g., ordinary least squares regression). Where there is spatial nonstationarity in the coefficients produced by a regression analysis, researchers using these global techniques may miss important local caveats in their predictions. When fully analyzed, these caveats can help to create better statistical models that might help to focus community resources and public policies in more effective ways.