Why have some of the findings on the democratic peace been divergent? Even within a single research program there may be serious differences in conclusions, based on the assumptions made by researchers and the methods used, Scholars who use the behavioral approach themselves point to some of the difficulties. Some researchers analyzing the democratic peace use different definitions of the key variables—democracy and war; for example, some researchers distinguish between liberal democracies (for example, the United States and Germany) and illiberal democracies (Yugoslavia in the late 19905). Also, the data for war would be different if wars with less than 1,000 deaths were included, as they are in some studies. And other studies of the democratic peace examine different time periods. Such differences in research protocol might well lead to different research findings. Yet even with these qualifications, the basic finding from the research is that democracies do not engage in militarized disputes against each other. That finding is statistically significant; that is, it does not occur by random chance. Overall, democracies are not more pacific than nondemocracies; democracies just do not fight each other. In fact, autocracies are just as peaceful with each other as are democracies, so that one could also talk of an autocratic peace.