A sprawling study of errors in decision-making, some literal matters of life and death.
You go to a doctor complaining of chest pains. The doctor orders an angiogram. The hospital requires a second opinion before authorizing surgery, and the second doctor disagrees on the extent to which a specific blood vessel is blocked. These unpredictable disagreements over the same data are what Kahneman, Sibony, and Sunstein call “noise,” a species of human error that happens whenever such higher-order judgments are involved. Noise, they write, is rampant in medicine, where “different doctors make different judgments about whether patients have skin cancer, breast cancer, tuberculosis, pneumonia, depression, and a host of other conditions.” Noise is especially prevalent in psychiatry, they add, where subjective opinion is more pronounced than in other disciplines. A cousin of bias, noise is difficult to isolate and correct. In forensic science, the authors write, noise is implicated in nearly half of all misidentifications of perpetrators and wrongful imprisonments. Unlike some categories of error, noise is often not helped by the introduction of more information. Writing in often dense but generally nontechnical prose, the authors offer strategies for reducing noise. One is to average out predictions in, say, stock market performance, since “noise is inherently statistical.” Another is to consult the smartest people you can find; while they may not be flawless, “picking those with highest mental ability makes a lot of sense.” Since error combines with snap decisions, the authors endorse rigorous review and other strategies for noise reduction and “decision hygiene” as well as developing habits of mind that acknowledge both bias and error and favor examining the opinions of those with whom one disagrees as dispassionately and fairly as possible. “To improve the quality of our judgments,” they urge, “we need to overcome noise as well as bias.”
Abundant food for thought for professionals of all types as well as students of decision science and behavioral economics.