The Book of Why

The term “causation” has long been banished from the lexicons of mainstream statisticians. The same taboo has made it difficult for scientists to address causal questions of great importance to society.

Humans naturally think in terms of cause and effect. For example, a three-year-old knows that tipping a cup will spill milk across the table. No mathematician will succeed in convincing the child that causes don’t exist and that only probabilities, relationships and correlations do. But when the child grows up and enrolls in higher-level math and statistics courses, he or she learns the heresy of causation as well as the mantra that “correlation does not imply causation.” Scholars become hesitant even to think about causation, let alone speak the word in published work. And yet, as normally functioning human adults, they reason causally all the time. For example, they press the brake pedal in full expectation that this will cause their cars to stop.

As late as the 1800s, scientists thought only in terms of causation. However, in the late 19th century, in an attempt to show causation in genetically inherited traits (that is, intelligence running in...