Authors: Shiloh Deitz*, University of Oregon
Topics: Spatial Analysis & Modeling, Water Resources and Hydrology, Population Geography
Keywords: census microdata, geospatial analysis, United States, demography, geographically weighted regression
Session Type: Paper
Start / End Time: 5:00 PM / 6:40 PM
Room: Balcony A, Marriott, Mezzanine Level
Presentation File: No File Uploaded
Undeniably, human phenomena – including inequalities – are spatially varying and exhibit spatial clustering. Geographers have made various methodological innovations contributing to our ability to capture these spatial patterns. In recent years, geographically weighted regression (GWR) has gained traction. However, due to the nature of human data (namely privacy and anonymity concerns) conducting GWR requires aggregate data. Scholars have largely taken the results of these analyses as proxies for household level patterns. However, do these results really approximate household or individual level patterns or are we committing the ecological fallacy? Drawing on census microdata in the United States, this study analyzes spatial patterns in demographic predictors of household level water security. I compare the results of a new method for measuring geographic patterns in household level regressions at the Public Use Microdata Area (PUMA) level to the results of GWR and synthesize and visualize the strengths and weaknesses of each methodology. I also explore ways that the multiple PUMA regression method and GWR might complement one another at various scales. Inequality is growing, and we are equipped with increasingly detailed geospatial human datasets; however, these require a conscientiousness to how the data are being used and what conclusions can or cannot be drawn. This study takes a modest step forward in advancing our methodological toolkit for exploring geographic variability in household level inequalities.