What to do (and not do) with foreign interventions data: Methodological challenges and solutions

The determinants of foreign interventions, be they military interventions or other forms of foreign involvement, are important for modern international relations and conflict studies. Nevertheless, major methodological problems with the analysis are often overlooked. This paper proposes modeling interventions as a two-stage decision-making process. The first stage determines whether a specific target state is intervenable from the perspective of a potential intervener, while the actual intervention target is selected among intervenable targets at the second stage. On this basis, we propose two solutions that allow obtaining estimates for predictors of an intervention: a statistical model that explicitly accounts for the two-stage nature of the data-generating process, and near-misses sampling. The two approaches are similar in that they treat intervention as a two-stage decision-making process. The authors utilize the second approach to create a sample of non-case observations. Two methods of data collection are used: matching and extracting information from open sources (media articles, reports, official documents). The authors conduct correlation analysis for several sample construction techniques to evaluate differences in empirical significance. The implications of the application of different sampling methods for the study of interventions are discussed.