The Questions of Causation

Causes and effects are ubiquitous in what society asks of scientists.

Will requiring students to take algebra improve test scores? Did global warming cause this heat wave? Does smoking cause cancer?

Statisticians carefully tiptoe around these questions. In Statistics 1 class, I taught my students the mantra “Correlation is not causation.” For example, countries with higher chocolate consumption tend to have more Nobel Prizes per capita. They’re correlated. But no one thinks that we can advance science by eating more chocolate.

Strangely, the textbook never explained causation. It never told the students how to diagnose and quantify causal effects, how to distinguish spurious from real. It failed them, and I failed them.

For nearly 30 years, my co-author, Judea Pearl, has been working to illuminate this blind spot. Statisticians refuse to address questions of causation, yet these are of greatest importance.

I joined Judea to write this book because I feel that it is profoundly important in our era of “big data.” Yes, data are useful and important, but they need to be combined thoughtfully and purposefully with causal models. “You are smarter than your data” is our new mantra. I hope our book will wake up a few scientists and empower non-scientists to demand real answers—not just correlations.

Available now, The Book of Why: The New Science of Cause and Effect (Basic Books) is Dana Mackenzie ’79's eighth book.