Statistics and Causal Inference

TLDR

Causal inference has long challenged statistics, as correlation does not imply causation yet well‑designed experiments can yield valid causal conclusions. The study seeks to determine what a statistical model can reveal about causation. The authors employ the Holland and Rubin causal inference model to critique the views of philosophers, medical researchers, statisticians, econometricians, and causal modeling proponents. Keywords include causal model, philosophy, association, experiments, Mill's methods, causal effect, Koch's postulates, Hill's nine factors, Granger causality, path diagrams, probabilistic causality, and the article appears in the collections Teaching Simpson’s Paradox, Confounding, and Causal Inference.

Abstract

Abstract Problems involving causal inference have dogged at the heels of statistics since its earliest days. Correlation does not imply causation, and yet causal conclusions drawn from a carefully designed experiment are often valid. What can a statistical model say about causation? This question is addressed by using a particular model for causal inference (Holland and Rubin 1983; Rubin 1974) to critique the discussions of other writers on causation and causal inference. These include selected philosophers, medical researchers, statisticians, econometricians, and proponents of causal modeling. Key Words: Causal modelPhilosophyAssociationExperimentsMill's methodsCausal effectKoch's postulatesHill's nine factorsGranger causalityPath diagramsProbabilistic causality This article is part of the following collections: Teaching Simpson’s Paradox, Confounding, and Causal Inference