Mega‐Review: Causality Books Causal Analytics for Applied Risk Analysis by Louis Anthony Cox, Jr., Douglas A. Popken, and Richard X. Sun. Springer, International Series in Operations Research & Management Science, Vol. 270, 2018, $229, xxii+588.The Book of Why: The New Science of Cause and Effect by Judea Pearl and Dana Mackenzie. New York: Basic Books, 2018, $32, x+419.Causality: Models, Reasoning and Inference by Judea Pearl, Cambridge, UK: Cambridge University Press, 2000

Abstract

These three books (which I shall refer to as CA, BW, and C) describe a revolution from traditional thinking in how causality is established and used in science and in decision making. The era of massive computer processing capability to deal with "big data" brings new opportunities and challenges to risk analysis. The conceptual issue of causality in science has a long history. The recent advances discussed in these books should help us to distinguish association from causation, where management choices can reduce risk. Those of us in the field of risk analysis need to be cognizant of this "causal revolution." These three books, written from different perspectives and for different audiences, give both an overview of the change in the analytical thinking, and, for the reader willing to invest the time and effort, detailed instruction in the methodology and computer algorithms.Those readers with a background in statistics would be well advised to read BW before CA and C. While the books have much in common, CA and C will be of greater help to the specialist who wants to learn how to do the analysis, rather than simply seeking understanding of the ideas. But CA and C are densely packed with advanced concepts. Those already familiar with Bayesian nets (BNs) and Bayesian reasoning might go directly to these books, but even they will learn much from BW, especially on why traditional statistics has avoided using both models and what-if (counterfactual) questions about actions that might reduce risks.