Logically equivalent descriptions of the same state of the world produce different choices because System 1 reacts to words, not just outcomes. Kahneman and Tversky's framing effects show preferences are frame-bound, not reality-bound.
Classic demonstrations
- Asian disease: "200 saved" → risk-averse; "400 die" → risk-seeking (same outcomes)
- Surgery vs radiation: 90% survival vs 10% mortality — identical facts, opposite physician choices
- KEEP £20 vs LOSE £30 — same sure outcome, different uptake
- Cash discount vs credit surcharge — economically identical, emotionally different
- Organ donation: opt-in (~4%) vs opt-out (~86%) — default frames laziness of System 2
Schelling's tax exemption: "larger exemption for rich" vs "larger surcharge for childless rich" — contradictory moral intuitions from reference point.
Implications
Framing is not always manipulation — some frames are better (gallons-per-mile vs MPG; sunk tickets vs lost cash). Policy: choice architecture (Thaler/Sunstein nudge). Personal: reframe losses as costs, widen mental accounts.
Tied to prospect-theory — gain/loss framing switches risk attitude.