The Book of Why, Judea Pearl (2.0)
I can’t say that I enjoyed reading this book, though it
did bring up some thoughtful questions about how causes and effects are treated
in science, specifically by statisticians. Unfortunately, the author has such a
high opinion of his own work; the book read as largely a pat on his back. He
summarizes theme of the book as ‘you are smarter than your data, data doesn’t
understand causes’. While this is true, the act of determining causes is mostly
fraught with assumption. The author tends to prove his point by using
hindsight, which, by definition, has fewer assumptions. In the book, he unveils
his ‘simple’ way of writing causal diagrams (90% of which were simple
triangles) that clarify how these causes are involved with the effect. Given
that the book was touted as written for non-statisticians and non-computer
scientists, I wish he would have included more real life examples to prove his
points and less time spent on the math.
No comments:
Post a Comment