Example · part of Thinking inside frameworks
You are asked to choose between two programs to fight a disease. One has a certain outcome, the other is a gamble. In 1981 Tversky and Kahneman put this choice to people in two wordings. When the outcomes were described as lives saved, 72% chose the certain program. When the identical outcomes were described as lives lost, 78% chose the risky one.
The facts were the same in both versions. Only the description changed. This is the framing effect: the wording supplies a reference point, and the reference point steers a choice between logically equivalent options. The book makes a related claim about messages. The same content can land differently depending on the key a reader decodes it with, the idea behind its three truth regimes.
The finding is robust. The Many Labs project reran it in 36 samples with 6,344 participants. The reversal appeared in 31 of them, and none showed the opposite. The effect there was smaller than in 1981. A 2020 survey of 88,181 people in 47 countries found it again. A smaller effect fits the wider pattern of influence, not imprisonment. The practical answer is to show both wordings.
How solid is this? Robust: replicated in 31 of 36 Many Labs samples, with a smaller effect than in the 1981 original, and found again across 47 countries.
Read the full argument
- Thinking Inside Frameworks, section “What a framework does to a thought”
How it connects
- What it shows: Defaults, Framing effect, Influence, not imprisonment
- See also: Crime as beast, crime as virus, Framework rotation, Truth-compression regimes (preservation / inversion / orthogonality)
- Evidence: Frameworks That Shape the Thoughts Built Inside Them, The Framing of Decisions and the Psychology of Choice
- Where the book discusses it: Thinking Inside Frameworks
← Framing effect · Thinking inside frameworks (13 of 21) · Conceptual metaphor →
Explore all the ideas · Read the book in order · Open the interactive map