Place Matters: Using Spatial Mapping to Show How Variable Relationships Differ Across School Districts

• In Event: Investigating Program Impacts Using Multilevel Models

Abstract

“Place matters” in moderating variable relationships because there are differences related to social, cultural, political, educational, and financial factors that shape local contexts. This paper examines relationships between variables for school district demographic composition, teaching and financial contexts, and student behavior/academic performance variables aggregated at the district level. The purpose is to investigate how these relationships vary across 565 districts within a state using spatial mapping. Local R2’s from geographically weighted regressions (GWR) are mapped using a geographic information system (GIS) to demonstrate variation in relationships across districts. Results show the importance of allowing relationships to “vary” between local contexts and that using a global measure of variable relationships based on aggregated data fails to capture important local variation.

Authors

• Mark Hogrebe, Washington University in St. Louis

• William F. Tate, Washington University in St. Louis