Using Behavioral Economics to Examine Community Shelter Use Intention in a Kansas Sample

Abstract Tornadoes remain a dangerous and unpredictable hazard in the U.S., particularly for individuals with limited access to safe shelter. Community tornado shelters are intended to provide protection for households lacking adequate in-home options. Still, access alone does not guarantee use. This study applied an operant behavioral economic framework to examine how distance to a community shelter shapes protective action during a simulated tornado warning. In a discrete-choice demand task embedded in a statewide survey, 1,859 Kansas residents indicated whether they would wait, shelter at home, or travel to a community shelter during a tornado scenario. Consistent with a barrier effect, travel time required to reach the community shelter reduced the odds of leaving for a community shelter, even after adjusting for demographics, housing characteristics, and hazard-related factors. Residents of multifamily and manufactured homes were more likely to use community shelters than those in single-family homes, whereas having a basement substantially reduced the likelihood of leaving. Socioeconomic patterns were also evident: middle-income households were most likely to leave, and higher educational attainment was associated with greater tendency to shelter at home rather than wait for further notice. These findings emphasize distance as a quantifiable, modifiable barrier to tornado safety and demonstrate how behavioral economic methods can inform evidence-based standards for shelter siting, access, and communication in support of improved hazard resilience.