4
THE EVIDENCE
This section relies on the most commonly used research designs, statistical tests,
and data to investigate the historical relationship between democracy and peace. The first
set of statistical results largely conforms to the conventional wisdom. Across nearly two
centuries of observations, the dyadic democratic peace--or the claim that democratic
states are less likely to engage in militarized conflict, including war, with each other--
holds up very strongly. On the other hand, during this same period, evidence for a
monadic democratic peace—or the claim that democratic states are less likely to engage
in military conflict with all regime types--is mixed. In a subsequent series of statistical
tests, I illustrate how aggregating data obscures differences in the effects of regime type
on war within this long historical period. These tests will show how the democratic
peace literature still faces an important historical challenge. Why does conclusive
evidence for a democratic peace not emerge until after World War I?
Research design
This section probes the historical effects of democracy on conflict with two
research designs, each with a separate unit of analysis.
1
The first employs the standard
dyad-year. Each possible combination of a pair of two states in a given year comprises a
separate observation.
The onset of a new militarized interstate dispute is used to operationalize conflict
within the dyad. The data is taken from version 3.0 of the Militarized Interstate Dispute
data set (Ghosn and Palmer, 2003). A militarized interstate dispute refers to “historical
cases in which the threat, display or use of military force short of war by one member
1
I also ran a series of statistical tests using the directed dyad research design. The results were largely
similar to those that follow. They underscored the links between democracy and conflict in the nineteenth
century; and then between democracy and peace in the twentieth century.
Convention