Sage Publications (London, Thousand Oaks, CA and New Delhi)

This article examines some unique features of a subset of crises in the International Crisis Behavior (ICB) dataset, involving cases in which nations perceived themselves to be in foreign policy crises when their primary adversaries did not. These ‘one-sided crises ’ are expected to differ from normal or ‘two-sided crises ’ in a number of ways, especially in terms of the likelihood of crisis violence. The absence of a shared perception about the likelihood of future hostilities at the onset of the crisis is posited as the basis for the expectation of differences between the two types of crisis. The study presents a number of hypotheses about these differences and provides a brief empirical analysis. A multivariate analysis indi-cates that one-sided crises are less likely to escalate to violence even when the effects of other known predictors of crisis violence are introduced. The study provides a cautionary note to scholars who use the ICB data about the potential impact of including one-sided crises in analyses of crisis violence. The conclusion provides a brief comment about the theoretical significance of the findings