Counterfactuals, hypotheticals and potential responses: A philosophical examination of statistical causality.

Abstract

Statisticians have developed a variety of conceptions, frameworks and tools for causal inference. We study some of these from a philosophical angle, focusing in particular on two formal frameworks, “Potential re- sponses” (PR) and “Decision Theory” (DT), and their use in problems of inferring the “effects of causes”. Although PR is currently the predomi- nant methodology, it is argued that DT is preferable both philosophically and pragmatically.