Abstract
Regional differences in behavioral responses to COVID-19 policies are not fully explained by demographic, socioeconomic, political, or psychological factors. This study examined whether a simulation-derived Cognitive Hierarchy (CH) parameter provides complementary information for explaining COVID-19-related behavior across 48 U.S. states and the District of Columbia. CH parameters were estimated from Google Community Mobility Reports using an agent-based simulation of recursive reasoning and incorporated into empirical models with conventional regional characteristics and Big Five personality traits. The CH parameter was positively associated with voluntary mobility reduction and remained statistically significant across alternative specifications, heteroskedasticity-consistent estimation, and spatial robustness analyses. The CH parameter also provided explanatory information beyond conventional regional characteristics and Big Five personality traits. External validation using COVID-19 vaccination uptake showed that CH remained positively associated with vaccination after accounting for personality, policy, and political factors. These findings suggest that simulation-derived behavioral parameters can complement conventional empirical measures in explaining heterogeneous COVID-19-related behavior.