Big wars, little wars: Avoiding selection bias

Much international conflict research focuses exclusively on big wars. Yet we lack adequate theory to designate some wars as candidates for bigness without relying upon knowledge of postwar events. Such reliance introduces selection bias into much research on war. That bias may vitiate the significance of our observations. A rudimentary representation of the relationship between rationality and expected war costs is tested against data on European wars from 1815–1965. The results are consistent with expectations. The results highlight the possibility to distinguish big wars from little wars ex ante without a separate theory for each. These simple analyses in conjunction with recent research on war diffusion remind us of the possibility that so-called big wars may not be different in their underlying logic from lesser wars, but rather represent the manifestation of different values on a set of generally important variables. That is, they are truly different in quantity, but not in quality.