Income Inequality and Health: Expanding Our Understanding of State-Level Effects by Using a Geospatial Big Data Approach

Abstract

Merging geospatial analytics with big data approaches provides a mechanism for leveraging and maximizing uses of traditional survey data to further extant work in meaningful ways. This study examines the income inequality hypothesis, which proposes that ecological (summary-level) income inequality is harmful for population health. However, findings from extant work are inconsistent across health outcomes and levels of geography. We contribute to this debate by applying a big data geospatial approach to create three innovative measures that capture uniformity in income inequality across counties within U.S. states. Using data from the Behavioral Risk Factor Surveillance System and American Community Survey, we evaluate multilevel models of individuals within states to examine the ways that income inequality, operationalized as the Gini coefficient, and three spatial uniformity measures that capture the way income inequality is dispersed across space within states, are associated with several health outcomes. Specifically, the uniformity measures capture the extent to which (1) inequality is uniformly distributed spatially in states regardless of whether the level is high or low, (2) the extent to which states are more uniformly high in inequality across space, and (3) the extent to which they are more uniformly low in inequality. We conclude that state income inequality did not predict worse health across these outcomes (and indeed was associated with lower odds of depression and obesity). However, residents of states that have more uniformly high inequality across space are more likely to report below-average health, cardiovascular disease, difficulty concentrating, and that they have not sought care because it was too expensive. We conclude with a discussion of how a big data geospatial approach can further contribute to research on this and other public health topics where scholars primarily rely on traditional survey data.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 26 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.