Appendix 12.B On Making Causal Inferences3

12.B.1 Introduction

This appendix addresses the issue of causality. While my motivation for discussing this slippery topic comes from the well-known utterance “correlation does not imply causation,” the scope and coverage of this discussion goes far beyond simply debunking the existence of any unequivocal connection between association and causation.

To gain a full appreciation of the essentials of causality, one has to make a fairly small investment in learning some important terminology—an investment that should yield a sizable payoff in terms of one's ability to think critically. That is, one has to become familiar with the rudiments of experimental design, logical relations, conditional statements, and the concepts of necessary and sufficient conditions. Given this grounding, the search for a working definition of causality and the link between logical implications and causality becomes a straightforward matter. It is then a short step to examining correlation as it relates to causality. Finally, we look to developing a counterfactual framework for assessing causality, and all this subsequently leads to the topic of testing for causal relations via experimental and observational techniques.

While the principal focus of this book is statistics, it is the case that there exists a profound difference between statistical and causal concepts. Specifically:

a. Statistical analysis: Employs estimation and testing procedures to determine ...

Get Statistical Inference: A Short Course now with the O’Reilly learning platform.

O’Reilly members experience live online training, plus books, videos, and digital content from nearly 200 publishers.