Regression to the Mean

Francis Galton described the pattern in the nineteenth century, and Daniel Kahneman met it in a lesson he never forgot. Teaching flight instructors in the Israeli Air Force, he argued that rewarding good performance works better than punishing mistakes. An experienced instructor objected: when he praised a cadet for a clean manoeuvre, the next attempt was usually worse, and when he shouted at a cadet after a poor one, the next was usually better. Kahneman saw that the instructor was watching regression. Every performance is part skill and part luck. An exceptional attempt was probably helped by luck, so the next one is likely to be closer to the cadet's average whatever the instructor says; a terrible attempt was probably unlucky, so the next will tend to improve anyway. Because praise tends to follow the best attempts and criticism the worst, the world seems to reward criticism and punish praise. The mind resists chance as an explanation. System 1 wants a cause, so regression gets credited to whatever happened in between: the shouting, the new manager, the change of coach, the treatment. Before you credit a change with an improvement, ask whether the starting point was extreme.