Framing Effects

Definition

Invariance — the idea that "two versions of a choice problem that are recognized to be equivalent when shown together should elicit the same preference even when shown separately" — is a bedrock assumption of rational choice theory. Framing effects are the demonstrated, systematic failure of this assumption: how a problem's outcomes are worded (as gains or as losses, relative to which reference point) changes the choice itself, not just how people talk about it. Kahneman and Tversky write that "in their stubborn appeal, framing effects resemble perceptual illusions more than computational errors" — even sophisticated respondents who are shown the contradiction and understand it cannot make it go away.

In the Book

The book's signature demonstration is the Asian disease problem. Told that a disease will kill 600 people, respondents overwhelmingly favor a certain option ("200 people will be saved") over an equivalent-expected-value gamble when the outcomes are framed as lives saved (72% choose the sure option) — but the majority reverses to favor the gamble when the identical outcomes are reframed as lives lost ("400 people will die"), with 78% now choosing the gamble. The two problems are mathematically the same choice; only the reference point (600 deaths as baseline versus zero deaths as baseline) changed. The book extends this with a dominance violation: splitting one framing-driven pair of decisions into two separately-presented decisions gets a majority of people to choose a combination that is provably worse than the alternative combination they rejected. Elsewhere in the book, McNeil, Pauker, Sox, and Tversky's study of physicians and patients choosing between cancer therapies described in terms of "mortality" versus "survival" statistics shows the same reversal in a real clinical decision, and the credit-card industry's insistence on "cash discount" instead of "credit card surcharge" wording shows framing exploited deliberately.

Why It Matters

Framing effects mean that presenting complete, accurate information is not sufficient to guarantee a stable decision — the same facts, worded differently, produce different choices, and neither wording is more "true" than the other. This has direct consequences anywhere someone controls how options are presented — policy communication, medical consent, pricing, negotiation — because framing can move outcomes without withholding or distorting any information, making it a lever that operates entirely legally and often unconsciously. The book's own prescription is to make framing effects visible by deliberately re-describing a decision in more than one way before trusting a strong preference.