Peace and Conflict Studies - Spring 2014

Peace and Conflict Studies Volume 21, Number 1 94 Graph I: Marginal effect of economic development in the probability of a MID in a developing- state dyad. Robustness Tests. The results of statistical analyses are sensitive to choices regarding variables, sample, and measurement. To test the robustness of my findings above, I reran my original model with different dependent variables, different measurements of the explanatory variables, and within different sub-samples. Table III displays these six replication models. Model 1 in Table III is a replication of the original model, which treated the onset and continuation of MIDs as the same, with onset MIDs only. Model 2 is a replication with a severer and more specific dependent variable: MIDs in which use of force materialized. Not all MIDs have equal seriousness and violence. Some remain as mere threats, some include actual use of force, and some escalate to full-scale wars. The dependent variable - useforce- equaled 1 if a dyad in a given year had a MID in which military force was actually used, and 0 otherwise. Model 3 replicates the original model within a sample that includes developing countries only. Dyads with a developed country are excluded. Model 4 is a replication of the original model within a smaller sample, which excludes dyads that include two of the six Middle Eastern oil-rich states, namely Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, and the United Arab Emirates. Oil-rich countries are rich but mostly not industrialized and therefore might deserve special attention when testing “capitalist peace” arguments. Model 5 is a replication with a dichotomous measure of democracy. A group of scholars argue that democracy should 0 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 GDP per capita in constant US dollars (1996, thousands) Probability of a MID (%)