Stochastic Variation: A Comment on Lewis-Beck and Skalaban's “The R-Squared”

In an interesting and provocative article, Michael Lewis-Beck and Andrew Skalaban make an important contribution by emphasizing several philosophical issues in political methodology that have received too little attention from methodologists and quantitative researchers. These issues involve the role of systematic, and especially stochastic, variation in statistical models. After briefly discussing a few points of disagreement, hoping to reduce them to points of clarification, I turn to the philosophical issues. Examples with real data follow.