The Book of Why: The New Science of Cause and Effect 1st Edition, Kindle Edition

The Book of Why is a popular introduction to Judea Pearl’s branch of causal inference. But it is also so much more.

Pearl has written many other textbooks introducing his graphical approach. But in this book, Pearl provides an engaging narrative of the history of causal inference, the important distinctions he sees in his branch and its importance for the future of Artificial Intelligence.

Briefly, Pearl views classical statistics as seriously flawed in not having developed a meaningful theory of causality. While able to demonstrate correlation, Pearl asserts that in classical statistics all relationships are two-way: that is 2x=3y+6 can also be written 3y=2x-6. We are left in doubt as to whether x causes y or y causes x.

Fundamentally, Pearl sees this problem as still plaguing all artificial intelligence and statistics. In its place, Pearl argues that the exact causal relationship between all variables should be explicitly symbolized in graphical form and only then can mathematical operations tease out the precise causal effect.

To be transparent, I am trained in the Rubin approach to causal inference and disagree with some of Pearl’s history and characterization of statistics. But that is not the point. The history is well-written, engaging and understandable by the lay reader. Similarly, his account of graphical causal inference theory is followable even for someone like myself who did not learn these techniques in graduate school.

The last part of the book, where Pearl opines on the future of AI, is the most sensational. Pearl believes that if computers were programmed to understand his symbolization of causal inference theory they would be empowered to realize counterfactuals and thus engage in moral decision making. Furthermore, since Pearl himself was a pioneer in deep learning, his characterization of contemporary AI as hopelessly doomed in the quest to replicate human cognition because of a lack of understanding in causal inference will be sure to garner attention.

But one would be misguided to think that speculations about AI or mischaracterizations of other kinds of causal inference make this book any less of a classic. For the first time, Pearl has written a popular, interesting and provocative book describing his branch of causal inference theory—past, present and future.

This book is a must read then, not only for causal inference theorists, but more widely for those with any interest in contemporary developments in computer science, statistics or Artificial Intelligence. A book that, like Kahneman’s Thinking Fast and Slow, is a triumphant summary of a lifetime of work in scientific topics that have ramifications, not only for fellow scientists, but for all of humanity.

Wow! I am a physician epidemiologist with a doctorate in epidemiology and I teach computational epidemiology (with R) at UC Berkeley. I had the opportunity to study biostatistics from the best professors at UC Berkeley School of Public Health (Steve Selvin, Nicolas Jewell, Richard Brand, and many more). The field of causal inference was just beginning to take off with biostatisticians piloting the plane (Mark van der Laan, Nicolas Jewell, etc.). I avoided a rigorous study of causal inference but eventually came around after studying Bayesian networks for decision analysis (FYI: Pearl pioneered Bayesian networks). Judea Pearl's Bayesian networks and causal graphs connects the fields of statistics, epidemiology, decision and computer sciences in a profoundly elegant way. His work empowers and expands the potential of "big data." This is the first book written for the general public on this topic. It will have a **huge impact**. Causality and causal reasoning is at the core of everything we see, do, and imagine. He provides a graphical tool (causal graphs) for encoding expert knowledge (including community wisdom and experience). Anyone --- yes, anyone --- can learn the basics. For additional rigor, there are structural causal models (functional equations). I now consider it data science "malpractice" to design studies, analyze data, or adjust for confounders without using causal graphs. As he covers extensively in the history of causality, human brains are wired to resist new paradigms. Be intellectually wise and humble and read this book -- you will not regret it!

This book represents the collaboration of Judea Pearl, a professional expert on causality, and Dana MacKenzie, an excellent science writer who is also a mathematician. It addresses the complicated differences between attempting to truly establish that A causes B versus erroneously asserting that A and B must be cause and effect because they correlate. The book is not a quick read. But the subject is of sufficient importance to merit the effort to digest its complex analyses. The section in the book treating the subject of climate change illustrates some of the challenges facing those seeking to truly understand what we know and don't know about this issue. Other examples are drawn from the field of medical research and other disciplines. A useful book written with great intelligence.