Andreas Wimmer and Brian Min

International Interactions, November 2009

Abstract

A critical question in the quantitative study of war is how to choose appropriate units of analysis. While most studies link wars to the sovereign states that fight them, several authors have recently offered alternative perspectives, focusing on the the properties of rebel organizations and armies or tracking events in fine-grained geographic spaces. We contribute to these developments by introducing a new dataset of wars fought from 1816–2001 on fixed territorial units of observation that conform to the grid of states in 2001. Compared to standard datasets, we expand the geographic purview to include states not recognized by the international system. We provide location codes that identify the territories on which conflicts are fought, which is especially important for the analysis of imperial wars and colonial rebellions. We also introduce a new typology of wars based upon the aims of warring parties rather than their status in the state system. This dataset is uniquely suited to explore new questions that cannot be addressed with other datasets. To illustrate, we test an institutionalist theory of war and show empirically that the types of wars fought in a territory depend on whether it is governed as a modern nation-state, an imperial dependency, or the center of an empire.