Relevance as a Latent Variable in Dyadic Analysis of Conflict

This study introduces split population binary choice models to address the issue of irrelevant dyads in the dyadic analysis of conflict with binary dependent variables. The advantage of employing a statistical model instead of directly identifying relevant dyads manifests itself in the selection of relevant dyads: rather than researchers making take-it-or-leave-it decisions, covariates are used to estimate relevance as a latent variable. An application of the model to the trade conflict debate shows that the unconditional probability of conflict, the quantity of interest for traditional binary choice models based on all dyads, is nonmonotonic with respect to trade when generated by the split population model. This finding provides one explanation for why trade is found either to increase or decrease the probability of conflict in existing research, as monotonicity by chance is imposed on an underlying nonmonotonic relationship.