| Abstract: | We argue that the prevailing dyad-year research design employed in quantitative tests of the theoretical literature on the democratic peace suffers from several limitations. We propose an alternative research design that focuses on three different stages or "games" that typically emerge in the evolution of international disputes over time. In this type of research design the units of observation are the policy choices of "challenger" and "target" states in each of these three games. We argue that this alternative research design has several advantages, such as requiring researchers to test a more complete and extensive set of hypotheses as well as helping them to address statistical issues related to strategic behavior and selection effects. |
