Causality from the Point of View of Classical Statistics

An attempt is made to provide a clear and concise basis for a statistical approach to causality which subsumes and reconciles the models proposed by J. Pearl, J. Robins, D. Rubin and other authors, and which fits in with classical statistical theory and with methods based on stratification and matching. Proofs of the most important results are given, and a variety of examples considered by the different schools of 'statistical causality' are treated in detail and in a self-contained manner.