Reconsidering Democracy and International Conflict: Does Conceptualization and Measurement Matter?

Are findings on democracy and conflict a product of unexamined conceptualassumptions about the nature of democracy inherent in the way we measureit? It is rare to find an empirical international conflict piece that doesnot mention that democracies do not fight each other. Very few studiescontradict this proposition. At the same time, there is reason forcaution due to how democracy is commonly operationalized and the conceptualassumptions inherent in the dataStudies that test for the effect of democracy most often rely on the Politydataset. Other authors have raised significant reservations about itsuncritical adoption (Gleditsch and Ward 1997). Our central concern is thatstudies which see democracy as affecting conflict may be picking uprelative differences between countries on the interval scale used by politythat do not capture qualitative differences between democracy anddictatorship. Would such findings hold if we considered democracy not inrelative but absolute terms? If we treated as democracies, only thosecountries that meet minimal standards (and operationalized it as a binaryvariable) like the literature on democratic survival (Gasiorowski,1995,Przeworski et al. 2000, Bernhard, Nordstrom, and Reenock2001), would we seedifferences?Two areas where past studies have strongly differed on this issue are theliteratures on the effects of democratization on conflict and theperformance of democracies in warfare. Mansfield and Snyder (1995)maintain that democratizing countries are more prone to interstateconflicts, while others have challenged their operationalization and do notfind a negative effect (see Thompson and Tucker 1997). Similarly withregard to the "democratic triumph" thesis (Lake 1992), there isdisagreement as to whether democracies are more successful inwarfare. Reiter and Stam (2002) find that democracies do win more butDesch (2002, 2003) challenges their findings and strongly criticizes theircases.In this study, we retest several important empirical works on democracy andinternational conflict using a measure of democracy, based on Dahl'sminimum criteria for polyarchy (1971). We use this measure to test therobustness of works such as "The Classical Liberals Were Right: Democracy,Interdependence, and Conflict: 1950-1985," (Oneal and Russett1997). "Democracies and War" (Reiter and Stam 2002), and "IncompleteDemocratization and the Outbreak of Military Disputes" (Mansfield andSnyder 2002). Our purpose is to see if their results are an artifact ofthe relative measure of democracy in the polity data, or whether they holdwhen we think of democracy as needing to meet certain minimum conditions. ..PAT.-Conference Proceeding [ABSTRACT FROM AUTHOR]