Bayesian Thinking about Macrosociology

Abstract

One question dominates methodological debate in macrosociology: In explaining large‐scale social processes should we seek simple theories that apply under a range of conditions or complex theories that are tailored to specific cases? Complex explanations are justified by their explanatory power and theoretical appeal. Simple explanations are justified by their falsifiability and their capacity to sustain strong conclusions. I weigh the merits of simple and complex explanations, applying some basic ideas from Bayesian statistics. The Bayesian analysis provides support for a compromise position that is rigorously grounded in the fundamental rules of probability. These ideas are illustrated in an analysis of labor union growth in Sweden.