Appendix B: Choices, Values, and Frames

Type: Appendix B

Category: Psychology

Difficulty: Advanced

Reading time: 5 min

Core idea

This appendix reproduces the core of Kahneman and Tversky’s 1984 paper “Choices, Values, and Frames” — the paper that brought together framing effects, prospect theory, and the analysis of decision under uncertainty into a unified framework for the general audience of American Psychologist. It is one of the most cited papers in behavioral economics.

The paper’s central argument: the standard theory of rational choice is descriptively wrong and prescriptively incomplete. People do not have stable preferences that are elicited by questions about choice. They have preference constructions that are sensitive to framing, to the mode of elicitation, and to the reference point from which options are evaluated. A theory of rational choice must account for these constructions — not just prescribe how a fully rational agent would behave.

Why it matters

The theoretical synthesis

“Choices, Values, and Frames” synthesizes the two major threads of Kahneman and Tversky’s work:

1. Prospect theory (from their 1979 Econometrica paper): the formal model of how people evaluate risky prospects, with reference-dependence, loss aversion, and the S-shaped value function.

2. Framing effects (developed in the 1981 Science paper with the Asian disease problem): the empirical demonstration that logically equivalent descriptions produce different choices, violating the invariance principle of rational choice theory.

The synthesis: framing effects are predictable from prospect theory. The gain and loss frames activate different parts of the S-shaped value function, producing predictably different risk preferences. Framing is not random; it follows the structure of the value function.

Invariance and its violation

A cornerstone of rational choice theory is invariance: if two descriptions of a choice problem are logically equivalent, they should produce the same choice. Kahneman and Tversky document systematic violations of invariance — the same facts, framed differently, produce reliably different choices.

This is not a peripheral finding. Invariance is a foundational assumption of rational choice — one that practitioners of economics, law, and medicine have relied on. If equivalent descriptions produce different choices, then any analysis of “what people prefer” must account for how the preference was elicited, not just what was chosen.

Author’s argument: The practical implication is that the design of choice is not neutral. How options are framed, what reference point is implicit, what is treated as the status quo — all of these are design variables that powerfully influence choices without changing the logical content of the options. Those who design choices shape them.

Choices and welfare

A deeper implication of the framing sensitivity finding: if choices are sensitive to logically irrelevant features, then choosing cannot be a reliable guide to welfare. The standard welfare economics assumption — that revealed preferences reveal what makes people better off — requires that choices be stable and invariant. They are not.

This opens the welfare questions that the two-selves discussion in the main text explores: if choices are not stable guides to welfare, what is the right welfare measure? Experienced well-being? Remembered well-being? Reflective preferences? These remain active research questions.

Key takeaways

Key takeaways

• The 1984 paper synthesizes framing effects and prospect theory into a unified account of how reference points, value function shape, and description format determine choices.
• Invariance violation: equivalent descriptions produce reliably different choices — a violation of the foundational rational choice assumption that equivalent descriptions should produce equivalent choices.
• Framing effects are predicted by prospect theory: gain frames activate the upper (concave, risk-averse) portion of the S-curve; loss frames activate the lower (convex, risk-seeking) portion.
• The design of choice is not neutral: the frame, the stated reference point, and the default option all influence choices without changing their logical content.
• Welfare implications: if choices are sensitive to framing, revealed preference cannot be a reliable measure of welfare — opening the two-selves questions about what measure of well-being is appropriate.
• The paper's lasting influence: co-founder of behavioral economics, cited across economics, law, medicine, and public policy — the canonical statement of framing in decision theory.

Mental model

click to zoom

The synthesis in 'Choices, Values, and Frames' — how prospect theory explains framing effects and why this matters for rational choice theory.

Read it as: Prospect theory’s S-shaped value function predicts framing effects: gain frames activate risk aversion (concave portion), loss frames activate risk-seeking (convex portion). This produces the systematic violation of invariance documented empirically. The theoretical implication is that rational choice theory — which assumes equivalent descriptions produce equivalent choices — is descriptively wrong. The practical implication is that choice design matters: frames, defaults, and reference points are powerful levers, not neutral presentation choices.

Practical application

Nudge design and the ethics of framing

The recognition that choice architecture powerfully influences choices without changing their logical content is the foundation of the nudge approach to policy (Thaler and Sunstein). Defaults in pension enrollment, organ donation, and savings plans can dramatically increase adoption without changing any incentive. The ethical question: who decides the defaults, and how? Using this power for the benefit of the chooser (libertarian paternalism) is different from using it for the benefit of the designer.

Applications:

• Organ donation: changing from opt-in to opt-out default dramatically increases donation rates without any change in legal rights — pure choice architecture effect.
• Savings rates: automatic enrollment in retirement accounts at a default contribution rate substantially increases participation — reference point is the enrolled state, not the non-enrolled state.
• Medical consent: informed consent processes that use equivalent information in gain vs. loss frames produce different rates of treatment acceptance — a framing effect with direct medical ethics implications.

Example

A company wants to increase employee 401(k) enrollment. Current approach: employees receive information about the benefits and must opt in. Enrollment rate: 30%. Change: employees are automatically enrolled at a 3% contribution rate and must opt out to decline. New enrollment rate: 85%. No change in incentives, information, or legal rights — only the default changed. The opt-out default activates the loss framing of status quo bias (leaving the enrolled state feels like a loss), dramatically increasing enrollment. This is the “Choices, Values, and Frames” synthesis in practice: the value function’s asymmetry around the reference point (the new default) shapes the choice.

Related lessons

Related

Chapter 34: Frames and Reality Framing effects in detail — the Asian disease problem and the empirical evidence for invariance violations. Chapter 26: Prospect Theory The theoretical foundation — the S-shaped value function that predicts and explains framing effects. Framing Effects The full concept — how presentation of information shapes choices independently of logical content.