The author found that the degree of national societal threat preceding congressional elections from 1946 to 1992 was positively associated with the mean state percentage of people voting for Republican representatives, supporting a conventional threat-authoritarianism hypothesis. However, threat was positively associated with the mean state percentage of people voting for Republican representatives in conservative states but not in liberal states, and the conventional threat-authoritarianism link was entirely driven by the relation in conservative states. The author classified states with a composite measure (α = .92) on the basis of state ideological identification, religious fundamentalism, composite policy liberalism, Republican Party elite ideology, and Democratic Party elite ideology. These results offer support to an interactive threat-authoritarianism hypothesis derived from the authoritarian dynamic theory of K. Stenner (2005), which postulates that only authoritarian persons are activated to manifest authoritarian behavior in times of normative threat. Also, the author discusses potential alternative explanations on the basis of system justification, need for closure, and terror-management theories.