The path from correlation to causation, to be sure, is fraught with pitfalls: if a crowd forms at a train platform, then often a train soon arrives.16 But crowds don’t impel trains to roll in. Something else—a train schedule—creates the correlation between crowds and trains.
