Case
Abstract
This Methods Case details the challenges and possibilities in examining domestic sources of foreign policy, specifically emergency assistance allocated to developing countries in response to humanitarian situations. The article focuses on modeling two-stage decision-making processes whereby the first stage causes a selection bias that needs to be accounted for. First, we discuss the possibilities to measure and quantify different aspects and subdomains of foreign policy, as well as the news media agenda. Next, we review the different pitfalls and possible solutions to modeling the relationship between media attention and government spending in a way that allows us to make causal inferences. Finally, we discuss the different assumptions and requirements for selecting the right statistical model for a two-stage decision-making process and zoom in on the specificities and practical use of the Zero-inflated Poisson regression model. The goal is to explain two-stage models in a way that is accessible and only modestly technical.