a dyad share this institutional feature and are able to exchange reliable information regarding unfolding events. The free press measures are so closely associated with the likelihood of conflict that the democracy score of the focal state, the more powerful state in each dyad, is never significant when one or more of the press freedom variables are also included in the specification. The effect of a free press on the probability of a dyadic dispute is substantively important as well. The likelihood of a dispute declines from .100 per year for a "typical" pair of contiguous states to .038 when both countries have independent presses.
In keeping with the contention that this can be considered as a refinement and extension of the work on the liberal, or democratic, peace it is clear that the results of this analysis could also be interpreted in terms of the norms-versus-structure debate within the literature on the democratic peace. The theoretical argument for including the press freedom measure in the analysis is basically a structural one. A free press matters because it enables the structural constraints in a liberal political regime to function more effectively in preventing international conflict. The interaction of free presses creates a decision-making context that raises the domestic political costs and decreases the leader's expectation of obtaining benefits from militarized conflict between free press countries. If the norms argument is correct, a free press is unnecessary because leaders should already be socialized against conflict in general and conflict with other democracies in particular. The statistical robustness of the free press measure as it is used in this analysis provides some reasonably strong support for the structuralist position although it clearly is not conclusive.