16 - Causes Trump Statistics

Описание: Causal explanations usually prevail over statistics. We prefer to believe in a concrete cause than to accept the role of chance or regression to the mean. The chapter shows how convincing stories eclipse data, affecting diagnoses, evaluations, and predictions.

Kahneman explores the fundamental tension between two ways of understanding the world: statistical thinking (based on base rates, probabilities, and population data) and causal thinking (based on stories, mechanisms, and individual cases). System 1 naturally generates causal explanations and is uncomfortable with statistical abstractions, which leads to systematic neglect of statistical information when it conflicts with compelling causal stories.

The chapter demonstrates that even when statistical information should dominate judgment, people give it insufficient weight if a causal story is available. In one experiment, participants received information about a cab involved in a hit-and-run accident. Some were told that 85% of cabs in the city are Green and 15% are Blue, and a witness identified it as Blue (but the witness is only 80% reliable). Others received the same witness testimony but were told that 85% of cab accidents involve Green cabs and 15% involve Blue cabs. Normatively, both versions should yield the same answer (about 41% probability that the cab was Blue). However, people gave much more weight to the statistical information when it was expressed as accident rates (a causal property) than when it was expressed as the composition of cabs in the city (a statistical base rate).

Kahneman explains that causal base rates are more readily integrated into judgment than statistical base rates. When base rates can be interpreted as propensities (dispositions to behave in particular ways), they become psychologically causal and therefore more intuitive. For example, knowing that "Green cab drivers are involved in 85% of accidents" suggests something about Green cab drivers (perhaps they're more reckless), making it feel relevant to judging an individual case. In contrast, knowing that "85% of cabs are Green" feels like mere bookkeeping that doesn't explain anything.

The chapter reveals how this preference for causal stories over statistics affects professional judgment. Medical students neglect base rates when diagnosing patients because they focus on symptom patterns. Hiring managers ignore selection ratios when evaluating candidates because they trust their assessment of the individual. Investors disregard market statistics when they have a compelling story about a company. In each case, specific information drives out statistical information, even when the statistics are highly diagnostic.

Kahneman discusses the psychological basis for this bias: System 1 is designed to construct causal narratives from observations, not to process abstract statistical regularities. Throughout human evolution, understanding cause and effect (if I do X, Y will happen) was more immediately useful than understanding probabilistic relationships. This legacy means our intuitive thinking naturally seeks and accepts causal explanations while resisting statistical reasoning.

The practical lesson is that statistical facts need to be embedded in causal stories to be psychologically compelling. To make base rates influential in judgment, express them as statements about individuals rather than populations: instead of "15% of students in this program fail," say "students in this program have a 15% failure propensity." This framing activates causal thinking and makes the statistical information more psychologically available. Understanding why causes trump statistics helps explain many judgment errors and provides strategies for making statistical information more influential in both personal and organizational decision-making.