The main purpose of this project is to test a new explanation of the process by which states go to war with one another. This explanation, which has been published and elaborated previously, is tested systematically with the new Militarized Interstate Dispute data of the Correlates of War project and the updated crisis data from the International Crisis Behavior project. These data, however, do not meet always and fully the needs of the project, especially with regard to territorial and non-territorial issues under contention. Therefore, the project requires relability and validity tests of data, cleaning of certain data, and, where necessary, new data collection.
After testing propositions derived from the explanation, it is appraised and then reformulated and made more precise, as appropriate. The steps to war explanation assumes that there are several paths by which war breaks out. It attempts to delineate the typical path by which the strongest states in the modern global system come to war. It conceptualizes the war process as emerging out of the interactions (and foreign policy practices) of states, with global systemic factors and domestic political factors playing important ancillary roles. It is claimed that war between equals is brought about by each side taking a series of steps that increasie hostility and make the issue(s) at hand more intractable. This involves disputants in a series of crises, one of which escalates to war. The published explanation specifies which steps are thought to increase the probability of war and which characteristics of crises (and disputes) make them prone to escalation. This project systematically tests these specifications with particular emphases on the role of territorial (versus non-territorial) disputes, the making of alliances, military buildups, repetition of disputes, and global norms. To investigate these specifications, we employ maximum likelihood estimation techniques suitable for the analyses of binary, ordinal, nominal, and duration level dependent variables. In addition, since some of our propositions imply reciproval causation and, thus, a need for more dynamic forms of statistical analyses, we utilize simultaneous equation models as well. Finally, we use two-step estimation procedures designed to model the potential presence of sample selection effects.
The project is the first systematic test of a new explanation of war and, as such, it expans the theoretical foundation of scientific work. It seeks to find the "causes" of war in something other than power and power distributions, which are the emphasis of most political science explanations of war. By focusing on something other than "power," the project hopes to increase scientific knowledge about other important political variables and their role in bringing about war. Testing the steps to war explanation increases the knowledge base of the field on the impact of foreign policy practices, especially with regard to how various steps and combinations of steps increase the probability of intense conflict. The project also hopes to make a significant contribution to the quality of data in the field on the types of revisions sought (territorial, policy, regime, miscellaneous) by challengers in militarized disputes.
