• Choose a category
• All
• Classic-Fiction

By Root 918 0

frequently resolve their conflicts with one another without resorting to war.123 The alternative of trying to select “non-crisis” as well as crisis cases is obviously problematic, and crisis cases have the advantage of posing tough tests for hypotheses supporting the democratic peace. Moreover, Elman selects cases that provide substantial variance on the independent and dependent variables, and in contrast to many studies, includes cases of both wars and successful crisis management from all three types of dyads: democratic dyads, mixed dyads, and nondemocratic dyads.124 Elman also evaluates alternative hypotheses via process-tracing and cross-case comparison tests.

The democratic peace literature has also produced commendable examples of research designs that incorporate “least similar” and “most similar” cases. In the most similar case design, the researcher attempts to select cases that are similar in all of their independent variables except one and differ in their dependent variable. James Lee Ray uses this design to compare the cases of the Fashoda Crisis and the Spanish-American War.125 Ray carefully addresses each of the standard categories of confounding variables identified by Donald Campbell and Julian Stanley, including “regression” (selection of cases with extreme scores on the variables of interest) and other kinds of selection bias, as well as “mortality” (the differential loss of respondents from the study: in this instance, the possibility that states may become more authoritarian as they see a war coming).126 Ray obviates other standard problems, such as the effects of history, maturation, and changes in instrumentation, by selecting cases from the same year. Ray also addresses six other variables that might account for the different outcomes of the two cases: proximity, power ratios, alliances, levels of economic development, militarization, and political stability. Ray’s systematic attention to these aspects of cross-case comparison, as well as his use of process-tracing evidence, bolsters his conclusions that Spain’s autocracy contributed to the Spanish-American War and that democracy in France and Britain helped peacefully resolve the Fashoda Crisis.127

As for the research design of least similar cases, Carol and Melvin Ember and Bruce Russett use the logic of this design, together with the instruments of statistical research, to test assertions about the democratic peace. In a least similar cases design, the researcher selects cases that are dissimilar in all but one independent variable, but that share the same dependent variable. This can provide evidence that the single common independent variable, in this instance the democratic decision-making processes, may account for the common dependent variable. In their study, Ember, Ember, and Russett test the findings of studies on the democratic peace among modern states and test them against pre-industrial societies; they find support for the proposition that participatory decision-making processes are conducive to peace in otherwise very different industrial and pre-industrial societies.

Both the most similar and least similar designs for case study comparisons, which rely on the logic of John Stuart Mill’s “method of difference” and “method of agreement,” respectively, are subject to methodological limitations that Mill and others have identified.128 In particular, the omission of relevant variables can entirely invalidate the results of cross-case comparisons in either design. Yet there are safeguards against this, as exemplified by Ray’s careful attention to a wide range of alternative hypotheses to ensure that no relevant variables are omitted from the comparison. In addition, process-tracing (undertaken by Ray, by Ember, and by Russett) provides an additional check on the results of cross-case comparisons.

Critiques and Challenges of Case Study Methods as Applied to the Democratic Peace

The case study literature on the democratic peace reveals two problems in case study methods: the problem of case selection and that of reconciling conflicting