There is an overwhelming amount of empirical evidence that democracy makes states more pacific toward their citizens. This robust finding has left scholars working in this area confident that they know {\em why} democracy causes states to be more pacific. I argue this is not true for two reasons. First, the theories adopted to explain this relationship have not been properly tested. Second, when good faith efforts have been made to test theories, measurement of all key variables has not been treated rigorously. I solve both of these problems by revisiting the theories upon which the literature rests and using a rigorous measurement strategy that is as true as possible to the theories proposed. I show that while the theories are up to the task of explaining the relationship, often the data are the weak link. Often, there is relatively little variation on the dependent and key independent variables. Thus, I show that most of the results generated in the literature are of the between-country variety rather than the within-country variety.