In this article, we seek to advance the theoretic and empirical literature on the diversionary use of force. We argue that state leaders are more likely to engage in diversionary foreign policy behavior when opposition groups from within the winning coalition press for policy changes, but the government rejects those policy demands. Only when domestic unrest threatens a loss of political support from groups that are politically important to the leadership do we expect leaders to try and rally their support through heightened international conflict. We test this argument in an analysis of Japanese foreign policy behavior from 1890 to 1941 regarding (i) the initiation of military threats, (ii) concessions in negotiations, and (iii) the escalation of military confrontations. Drawing upon new data sets collected on Japanese domestic politics and foreign relations, we find strong support for our argument in a series of statistical tests.There is an extensive scholarly literature on the relationship between domestic and international conflict. In this paper, we focus on the ''diversionary use of force'' literature in which the central question is whether political leaders experiencing internal political turmoil are more likely to engage in confrontational foreign policy behavior. A common claim is that during such periods of domestic unrest, government leaders attempt to divert public and elite attention away from internal problems by initiating and escalating international conflicts. The argument is that leaders anticipate rallying domestic support in opposition to an external threat, which should weaken prevailing domestic political discontent 1 Authors' notes: We thank the ISQ editors, the anonymous reviewers, and members of the CIDCM Workshop on Peace and Conflict for their helpful comments. Replication materials are available at Paul Huth's Dataverse International Studies Quarterly (2010) 54, 915-937