In this chapter, I explain the research design necessary to test the previously formulated hypotheses. I will first devote attention to the case selection (i.e., which armed groups are included in the dataset) and how I measure civilian killings. Thereafter, I explain in detail how I have collected systematically and empirically information on the internal structure of the different armed groups under scrutiny. In doing so, I explain how I have made the Web survey, who was selected to participate in this survey, and how I have measured the different independent variables.