SAR-Gi*: Taking A Spatial Approach to Understand Food Deserts and Food Swamps

Authors: He Jin*, Texas State University - San Marcos, Yongmei Lu , Texas State University
Topics: Food Systems, Spatial Analysis & Modeling, United States
Keywords: economic deprivation, sociocultural deprivation, geographic access to food outlets, food desert, food swamp
Session Type: Virtual Paper
Day: 4/10/2021
Start / End Time: 11:10 AM / 12:25 PM
Room: Virtual 31
Presentation File: No File Uploaded


Most existing assessment methods of food deserts and food swamps adopts a non-spatial approach, overlooking the role of spatial dependence in the associations between food access and socioeconomic status. This study aims to fill the gap by examining this relationship with a spatial perspective. Using Austin, TX as a case study, we adopted a multi-mode Huff-based 2SFCA method to measure the spatial access to food stores. Eight socio-demographic variables were represented by two indices per factor analysis: Economic Deprivation Index (EDI) and Sociocultural Deprivation Index (SDI). We proposed a SAR-Gi* model to characterize food deserts and food swamps in the Austin area at the census block group level. The SAR-Gi* model emphasizes the spatial dependence between spatial access to food stores, EDI, and SDI. Our analyses reveal that EDI is a significant predictor for access to healthy food (β = -0.054, p = 0.037), and SDI is significantly associated with access to unhealthy food (β = 0.160, p = 0.000). We also found that food deserts in Austin are concentrated in the eastern part of the city while food swamps in the northeast part. The noticeable difference between the spatial patterns of food deserts and food swamps as identified by our study and those based on the general USDA definition or other traditional statistics methods speaks to the great potential of the SAR-Gi* model in reflecting geographically specific patterns and relations embedded in food access and socioeconomic status of neighborhoods.

Abstract Information

This abstract is already part of a session. View the session here.

To access contact information login