Infectious disease outbreaks are expected to predict support for conservative policies. However, earlier studies (January–June, 2020) reached conflicting findings regarding the association between COVID-19 threat and policy preferences in the United States. We revisit this issue by analyzing five nationally representative surveys conducted during the relatively severe periods of the pandemic (August 2020–December, 2020; total N = 82,753). Using Bayesian inference, we find strong evidence that subjective (e.g., fear of infection and pandemic outrage) but not objective (e.g., local cases and deaths) threat predicted support for liberal policies (e.g., immigration and universal health care). Meta-analyses revealed that the estimates depend on the type of subjective (.05 ≥ r ≤ .60) or objective (.00 ≥ r ≤ .14) COVID-19 threat. We propose an emotion-mediated dual-process model of pathogen management suggesting that infectious disease outbreaks activate both avoidance and caregiving motives that translate, respectively, into support for right-wing and left-wing policies.