Case Studies and Theory Development in the Social Sciences - Alexander L. George [25]
This chapter looks at these three generations of the literature on the interdemocratic peace. Yet this tripartite categorization of research on this topic should not be taken as suggesting that any one method has or will supplant others in the democratic peace research program or that the evolution of social science research programs generally proceeds from one method to another. Research using all three methods usually proceeds simultaneously and iteratively, as each method confronts new research tasks where another method is superior. Much useful work on the democratic peace remains to be done using all three approaches. As case studies and formal models refine the concepts and logic of democratic peace theories, statistical tests can fruitfully be redone using these new concepts and their associated measurements. Such tests will in turn help identify new sets of anomalous cases for further case studies, which can provide fertile ground for both inductive and formal refinements to extant theories, which will need to be tested by new statistical studies, and so on.
The First Generation: Contributions of Statistical Methods
As James Lee Ray points out in his thorough review of the literature on the democratic peace, arguments for the existence of a democratic peace can be traced back to such liberal theorists as Immanuel Kant and Woodrow Wilson, and critiques of these arguments have an equally distinguished pedigree among realist thinkers like E. H. Carr and Hans Morgenthau and neorealists such as Kenneth Waltz.83 Much of the contemporary research on this subject, however, can be traced back to a 1964 article by Dean Babst, a research scientist at the New York State Narcotic Addiction Control Commission.84 In this four-page article, Babst concluded that “no wars have been fought between independent nations with elective governments between 1789 to 1941,” and he calculated the difference between the proportions of democratic and mixed or nondemocratic dyads at war in World Wars I and II to be significant at the 1 percent level.85 Yet Babst’s article was theoretically underdeveloped, positing a monadic explanation (the purported reluctance of democratic publics to vote to take on the costs of war) for this dyadic result, and it did not control for important variables.
J. David Singer and Melvin Small rescued Babst’s argument from obscurity among political scientists by critiquing it in a 1976 article that contended that the war involvement of democratic states between 1816 and 1965, in terms of duration and battle deaths, was not significantly different from that of other types of regimes. Singer and Small suggested that the absence of wars between democracies was due to the fact that democratic states rarely bordered upon one another, but they did not test this assertion.86 In the late 1970s and 1980s, a rapidly expanding body of statistical research made three key contributions to the democratic peace research program. First, statistical studies refined the research question from whether democratic states were more peaceful in general to whether they were more peaceful only or primarily toward one another (the interdemocratic peace).87 Some research continued on the monadic proposition that democracies might be more peaceful in general, but research increasingly focused on the stronger evidence for an interdemocratic peace.88 Researchers also used statistical methods to test whether democracies have been less likely than other states to engage in conflicts short of war—both generally and vis-à-vis one another.89 Some researchers also began to examine whether subtypes of states, such as states in transition to democracy, were more or less prone to war.90 Second, many statistical studies tested for whether findings of an interdemocratic peace were spurious by controlling for the effects of numerous variables—including contiguity, wealth, alliance membership, relative military capabilities, rates of economic growth, and the presence of a hegemon.91 Third, researchers using statistical methods theorized on and began