Scale and Local Modeling: New perspectives on voting behavior in the United States

Keywords: MAUP, Multiscale Geographically Weighted Regression, Local models, Operational scale of process
Session Type: Virtual Paper Abstract
Day: Friday
Session Start / End Time: 2/25/2022 02:00 PM (Eastern Time (US & Canada)) - 2/25/2022 03:20 PM (Eastern Time (US & Canada))
Room: Virtual 26

Abstract

The well-known Modifiable Areal Unit Problem (MAUP) has to date almost exclusively been investigated from the perspective of data properties using global models of spatial analyses. With the advent of local models that focus on potentially spatially heterogeneous processes, there is an opportunity to revisit the problems related to spatial scale from the perspective of processes. To explore how process properties might affect the sensitivity of global modeling results, both simulated and empirical experiments are conducted. In a recent paper using simulated data we confirm the initial hypothesis that the sensitivity to the MAUP in global spatial analysis is associated with the degree of process nonstationarity. The primary aim of this research is to test these initial findings using high quality granular voting data from the U.S. Presidential elections in 2016 and 2020 and exploring the sensitivity of the processes controlling voter preferences to variations in spatial scale.