A multilevel regression model to explain disparities in residential cooling infrastructure across heat-threatened regions

Abstract

Residential Cooling Infrastructure (RCI) plays a crucial role in mitigating heat stress among residents. However, the unequal distribution of RCI remains a largely overlooked dimension of thermal inequality, and large-scale, fine-resolution assessments of the multilevel drivers of this disparity are still scarce. This study identifies significant disparities in RCI prevalence, affecting at least 2.32 million residents in the state of Texas, USA. The results of our multilevel analysis reveal that the proportion of the population with low educational attainment is the most powerful and consistent predictor of low RCI prevalence at both the census tract and county levels. In particular, areas with higher RCI prevalence report average earnings 1.6 times greater than those with lower prevalence. A one-standard-deviation increase in this variable raises the odds of low RCI prevalence by up to 198% at the census tract level. High Intraclass Correlation Coefficient (ICC) values (0.53–0.62) further demonstrate that RCI inequity is not simply a localized segregation phenomenon but a systemic challenge, with over half its variance rooted in county-level factors. These findings underscore the critical role of education in enabling individuals and communities to adopt and maintain adequate cooling infrastructure, supporting the formulation of public policies that promote educational advancement as a feasible strategy to enhance societal resilience to heat stress.