Quantitative social science relies centrally on the ideal that any concept can be operationalized and measured. Although notionally tenable, practical limits dictate that some concepts might be more easily and more reliably measured than others. Some conceptual variables are not easily operationalized or measured because they are not directly observable phenomena; as a result, scholars often resort to indirect indicators of the phenomenon in question. Opportunity comprises one such concept that is important to the development of theories of international conflict. Although scholars might agree that different states have different opportunities to fight, measuring opportunity has proved to be a daunting task. This study presents a statistical method that treats unobserved concepts (such as opportunity) as latent variables. The procedure derives estimates of the latent variable and observable actions that arise from the latent process. Drawing on split population hazard models developed in the economics literature, a model of dyadic opportunity and willingness to engage in interstate conflict is produced. Results shed light on the process that leads states to fight one another.
