Stated Preference¶

Version: v3 · 2026-09-07 · History
Prime #: 1463
Origin domain: Economics & Finance
Subdomain: preference measurement → Economics & Finance
Also from: Survey Methodology, Psychology, Machine Learning
Aliases: Declared Preference, Expressed Preference, Elicited Preference, Self Report Preference
Related primes: Revealed Preference, Preference, Elicitation Channel Contribution

Core Idea¶

Stated preference is the evidence rule that infers an agent's valuation from what the agent reports when asked, rather than from what the agent chooses when real alternatives and consequences are present. The datum can be a ranking, survey response, declared willingness to pay, vote intention, policy preference, or model-generated statement of a desired policy.

The method earns its place because it reaches values that behavior cannot yet reveal. Analysts can ask about a public good with no market, a product that does not exist, a contingency the respondent has never faced, or an option excluded from the current menu. Revealed behavior is silent in those cases; a report is available.

The cost is that the report is usually cheap and channel-mediated. Nothing may be forgone by naming a preference, and the recorded answer is produced through wording, ordering, response options, social setting, incentives, and a respondent's ability to predict future behavior. The report can be sincere yet still differ from what the same agent would choose. Stated Preference therefore names an access mode to valuation, not a claim that the spoken answer is the valuation itself.

Structural Signature¶

The latent preference — an ordering over alternatives that the analyst seeks to recover.
The reporting agent — an entity able to state, rank, or otherwise communicate a valuation.
The elicitation act — a question, prompt, survey, interview, or stated-choice task.
The report — the recorded answer treated as evidence about the latent ordering.
The low-cost channel — reporting usually imposes less opportunity cost than acting.
The reach advantage — hypothetical, unavailable, public, or future alternatives can be valued.
The distortion profile — hypothetical bias, strategic response, social desirability, articulation limits, and channel contribution.
The validity test — change the channel or compare later behavior and observe whether the inferred ordering persists.

What It Is Not¶

Not Preference itself. Preference is the ordering; Stated Preference is one inference method for accessing it.
Not Revealed Preference. Revealed Preference credits observed choice under a real menu; Stated Preference credits a solicited report.
Not automatically false or insincere. A report can be sincere and still fail to predict action because the future choice context differs or the respondent cannot forecast it.
Not merely Cheap Talk. Strategic costless speech is one failure mode, but channel effects and hypothetical error occur without deception.
Not a survey format. Surveys are one implementation; interviews, stated-choice experiments, and model prompts instantiate the same report-as-evidence relation.

Broad Use¶

Environmental and public economics: contingent valuation where no market purchase reveals value.
Transportation and product design: stated-choice experiments over services or features not yet offered.
Polling and public policy: declared priorities, vote intention, and support for hypothetical policies.
Psychology and health: self-reported attitudes, intentions, treatment preferences, and risk tolerance.
User research and marketing: claimed willingness to buy, switch, subscribe, or adopt.
Machine learning: direct preference queries or stated policies compared with logged behavior.

Clarity¶

The abstraction forces a clean distinction among three objects that are often collapsed: the agent's underlying valuation, the report produced by an instrument, and the analyst's inference from that report. Once separated, disagreement with later behavior is not automatically hypocrisy. It may identify a channel effect, a hypothetical-versus-situated context change, a constrained menu, or a genuinely unstable preference.

Manages Complexity¶

Many instrument-specific disputes compress into one evidence-channel trade. The report channel has high reach and low behavioral cost; the choice channel has real stakes but limited menus and environmental confounds. An analyst can therefore choose the channel by what must be learned, then apply the appropriate validity checks instead of treating one source as universally superior.

Abstract Reasoning¶

Let \(P\) denote the latent preference, \(C_s\) the stated-preference elicitation channel, and \(R_s=f(P,C_s)\) the recorded report. Inference from \(R_s\) to \(P\) is an inverse problem. Identification improves when the analyst varies \(C_s\), introduces incentive-compatible consequences, checks internal consistency, or triangulates against choices produced under a documented menu and cost structure.

The decisive reasoning move is to ask what the report can reveal that behavior cannot, and what the channel can add that does not belong to the preference. That two-sided test prevents both reflexive dismissal of self-report and naive acceptance of it.

Knowledge Transfer¶

The same structure runs through contingent valuation, polling, patient decision aids, user research, preference learning, and model evaluation. In every case, an agent is queried, the answer is treated as evidence of valuation, inaccessible alternatives become measurable, and validity depends on the response function of the elicitation channel. The application vocabulary changes; the report-versus-choice evidence rule does not.

Examples¶

Nonmarket valuation¶

A city asks residents how much they would pay to preserve a wetland that has no market price. The responses expose a valuation unavailable from purchase records, but estimates can shift with the payment vehicle, scope description, and whether payment is credible. The method's reach and channel dependence appear together.

Product research¶

Users rank proposed features before any version exists. The ranking is valuable because no behavior can yet reveal demand. Later usage may diverge because implementation quality, switching costs, defaults, or real prices enter the choice. The earlier answer remains a stated preference for its elicitation context, not proof of future adoption.

Structural Tensions¶

Reach versus incentive: the ability to value the hypothetical is purchased with a signal that may carry little real cost.
Sincerity versus predictive validity: honest reports can fail to predict behavior.
Channel control versus ecological realism: highly controlled elicitation may remove the very context that shapes real choices.
Individual valuation versus constructed response: the instrument can help respondents form an answer rather than merely record a pre-existing one.
Symmetry with Revealed Preference: both are inference methods with different confounds; neither deserves automatic epistemic priority.

Structural–Framed Character¶

Stated Preference is a framed prime on the structural–framed spectrum: its meaning is bound to a methodology for valuing things, and it arrives with that methodology attached. It is the evidence rule that infers what an agent values from what the agent reports when asked, rather than from what the agent chooses when real alternatives and real consequences are on the table.

It cannot be defined without human practice: there must be someone asked, a question posed, and an answer given. With it comes an economics lexicon — willingness to pay, contingent valuation, elicitation, cheap talk, revealed preference — that stays intact wherever the prime is used, whether the material is a survey about a public good that has no market or a language model's declared policy choice. Applying it to either recasts the case as valuation elicited from a respondent. Its evaluative charge is the mildest part — the standing critique is methodological rather than moral — but on the remaining diagnostics it reads framed.

Structural Core vs. Domain Accent¶

The portable core is simple: query an evaluator and infer an ordering from the returned representation. Economics contributes willingness-to-pay, contingent valuation, and stated-choice terminology; psychology contributes self-report; polling contributes declared intention. Remove those accents and the report-based access mode remains.

Preference is the latent ordering the report is used to infer.
Elicitation Channel Contribution identifies the instrument's systematic share of the recorded answer.
Revealed Preference is the choice-based dual.
Stated–Revealed Preference Gap compares the two access modes when their outputs diverge.

Relationships to Other Abstractions¶

Current abstraction Stated Preference Prime

Parents (2) — more general patterns this builds on

Stated Preference is part of, typical Elicitation Channel Contribution Prime

A stated-preference estimate usually contains a systematic contribution from the question, prompt, scale, or protocol through which the report was elicited.
Stated Preference presupposes Preference Prime

Stated Preference presupposes an underlying preference because the solicited report is evidence about an ordering, not the ordering itself.

Children (1) — more specific cases that build on this

Stated–Revealed Preference Gap Prime is part of Stated Preference

The gap contains the report-based estimate whose output is compared with behavior under a real choice opportunity.

Hierarchy paths (3) — routes to 3 parentless roots

Stated Preference → Preference

Show alternative paths (2)

Neighborhood in Abstraction Space¶

Stated Preference sits in a moderately populated region (50^th percentile for distinctiveness): it has near-neighbors but no dense thicket of synonyms.

Family — Preference, Utility & Choice (18 primes)

Nearest neighbors

Stated–Revealed Preference Gap — 0.80
Revealed Preference — 0.76
Preference — 0.72
Temporal Inconsistency and Preference Reversals — 0.70
Salience-as-Significance — 0.69

Computed from structural-signature embeddings · 2026-09-10

Solution Archetypes¶

No catalogued solution archetypes reference this prime yet.