744 Power or Plenty? between 1918 and 2001.2 Brecher and Wilkenfeld (1997,4-5)explain that these events were selected based on two criteria: (1)a change in type and/or an increase in intensity of disruptive,that is,hostile verbal or physical interactions between two or more states,with a heightened probability of military hostilities;that,in turn,(2)destabilizes their relationship and challenges the structure of an international system-global,dominant,or subsystem. Although any crisis could affect American economic or security interests,not all the incidents included in the ICB data are relevant here.In order to test hypotheses about intervention realistically,the events examined must be reason- ably comparable.Decisions about intervention in an ongoing crisis differ from decisions to initiate a crisis in the first place.Bremer (1992),Bennett and Stam (2003),and others treat initiation of wars differently from intervention in ongoing wars for this reason.Although decisions to initiate crises certainly reveal something about intervention,the fact that there are only 20 such cases in the ICB data set limits the analytical usefulness of these events.Only decisions to join ongoing crises will be considered here.Second,some of the crises included in the ICB data occurred during wars.For example,24 crises were part of World War II,14 of them beginning after the United States had become a belligerent. Just as decisions about initiation differ from decisions about intervention,inter- vention in events taking place during wars in which the United States was already a belligerent also differ.These crises are also omitted in the analysis that follows. Finally,crises in which the United States was the target of actions by other states are not appropriate for testing hypotheses about intervention because American leaders were not able to make a decision about their involvement.Overall,393 of the 435 crises that began before 2002 were selected for analysis here. Like Regan's data,the ICB data record major power actions short of direct military intervention.The ICB data include information on economic aid,propa- ganda activity,covert action,and statements by major foreign policy officials.As with civil wars,it makes sense to set the threshold for intervention lower than full-fledged military action,which is likely to be a last resort after less costly policy instruments have failed.For purposes of the analysis presented here,the United States and its rivals were coded as intervening if they undertook some action beyond official statements and propaganda.In terms of the ICB data, intervention included military action,the provision of military advisors,military or economic aid,and covert action.By this standard,the United States inter- vened in 94 of the 393 crises selected for analysis. Operationalizing Economic and Security Interests Because security interests can encompass a wide range of considerations,includ- ing many that are difficult to observe,operationalizing the concept is not easy The analysis here will consider two considerations that have the virtue of being both important and easy to observe.The first is the presence of an alliance with one of the affected states.The security of allies has often been a major concern of American policy makers.For purposes of this analysis,states are considered allies if they have a defensive or offensive alliance with the United States accord- ing to the Alliance Treaty Obligations and Provisions (ATOP)data set (Leeds et al.2002).Civil wars taking place in American allies should be more likely to prompt U.S.intervention.Similarly,international crises in which an American 2 The ICB data include 10 additional crises from 2002 through 2004,but data on the major independent variables used here are not available for these years.between 1918 and 2001.2 Brecher and Wilkenfeld (1997, 4–5) explain that these events were selected based on two criteria: (1) a change in type and ⁄ or an increase in intensity of disruptive, that is, hostile verbal or physical interactions between two or more states, with a heightened probability of military hostilities; that, in turn, (2) destabilizes their relationship and challenges the structure of an international system—global, dominant, or subsystem. Although any crisis could affect American economic or security interests, not all the incidents included in the ICB data are relevant here. In order to test hypotheses about intervention realistically, the events examined must be reason￾ably comparable. Decisions about intervention in an ongoing crisis differ from decisions to initiate a crisis in the first place. Bremer (1992), Bennett and Stam (2003), and others treat initiation of wars differently from intervention in ongoing wars for this reason. Although decisions to initiate crises certainly reveal something about intervention, the fact that there are only 20 such cases in the ICB data set limits the analytical usefulness of these events. Only decisions to join ongoing crises will be considered here. Second, some of the crises included in the ICB data occurred during wars. For example, 24 crises were part of World War II, 14 of them beginning after the United States had become a belligerent. Just as decisions about initiation differ from decisions about intervention, inter￾vention in events taking place during wars in which the United States was already a belligerent also differ. These crises are also omitted in the analysis that follows. Finally, crises in which the United States was the target of actions by other states are not appropriate for testing hypotheses about intervention because American leaders were not able to make a decision about their involvement. Overall, 393 of the 435 crises that began before 2002 were selected for analysis here. Like Regan’s data, the ICB data record major power actions short of direct military intervention. The ICB data include information on economic aid, propa￾ganda activity, covert action, and statements by major foreign policy officials. As with civil wars, it makes sense to set the threshold for intervention lower than full-fledged military action, which is likely to be a last resort after less costly policy instruments have failed. For purposes of the analysis presented here, the United States and its rivals were coded as intervening if they undertook some action beyond official statements and propaganda. In terms of the ICB data, intervention included military action, the provision of military advisors, military or economic aid, and covert action. By this standard, the United States inter￾vened in 94 of the 393 crises selected for analysis. Operationalizing Economic and Security Interests Because security interests can encompass a wide range of considerations, includ￾ing many that are difficult to observe, operationalizing the concept is not easy. The analysis here will consider two considerations that have the virtue of being both important and easy to observe. The first is the presence of an alliance with one of the affected states. The security of allies has often been a major concern of American policy makers. For purposes of this analysis, states are considered allies if they have a defensive or offensive alliance with the United States accord￾ing to the Alliance Treaty Obligations and Provisions (ATOP) data set (Leeds et al. 2002). Civil wars taking place in American allies should be more likely to prompt U.S. intervention. Similarly, international crises in which an American 2 The ICB data include 10 additional crises from 2002 through 2004, but data on the major independent variables used here are not available for these years. 744 Power or Plenty?