Choice Experiment Valuation
Also known as: Discrete Choice Experiment, DCE, Choice-Based Conjoint Valuation, Stated Choice Experiment
A choice experiment (discrete choice experiment, DCE) is an attribute-based stated-preference method that values non-market goods by describing them as bundles of characteristics and asking respondents to choose repeatedly among competing alternatives — one of which always carries a cost. Grounded in random utility theory, the choices are modeled with a discrete-choice model whose coefficients reveal the relative value of each attribute, and dividing any attribute's coefficient by the cost coefficient yields its marginal willingness to pay.
Key highlights
- Recovers the marginal value of individual attributes and the trade-offs between them, not just a single holistic value.
- Naturally values new or non-existent goods and policy scenarios by recombining attribute levels.
- Choosing among options is cognitively easier and arguably less prone to some biases than naming a dollar amount.
- Built on a rich random-utility econometric framework (logit, mixed logit, latent class) enabling preference heterogeneity and welfare analysis.
Intuition
This section is available to Pro members. Upgrade to Pro
How it works
This section is available to Pro members. Upgrade to Pro
When to use it
Use a choice experiment when the good of interest can be characterized by several attributes and you want the marginal value of each attribute and the trade-offs among them — for example pricing the separate contributions of biodiversity, water quality, and access in an environmental program, or valuing features of a health intervention or transport option. It is the stated-preference method of choice when policy can vary attributes independently or when you want to value scenarios not yet observed. Like contingent valuation it relies on hypothetical responses and demands careful attribute definition, efficient experimental design, and tests for hypothetical and choice-set effects; when only a single holistic change (and especially a pure non-use value) is at issue, a contingent valuation referendum may be simpler and less prone to attribute-processing complications.
Strengths & limitations
- Recovers the marginal value of individual attributes and the trade-offs between them, not just a single holistic value.
- Naturally values new or non-existent goods and policy scenarios by recombining attribute levels.
- Choosing among options is cognitively easier and arguably less prone to some biases than naming a dollar amount.
- Built on a rich random-utility econometric framework (logit, mixed logit, latent class) enabling preference heterogeneity and welfare analysis.
- Still a stated-preference method, so subject to hypothetical bias between stated and actual behavior.
- Cognitive burden rises with the number of attributes, levels, and choice tasks, inducing simplifying heuristics and attribute non-attendance.
- Results depend on attribute and level selection and on the experimental design; omitted or poorly defined attributes bias estimates.
- The basic conditional logit imposes the independence-of-irrelevant-alternatives restriction, requiring richer models (mixed/nested logit) when it fails.
Common pitfalls
This section is available to Pro members. Upgrade to Pro
Applications
This section is available to Pro members. Upgrade to Pro
Frequently asked
How is marginal willingness to pay obtained from a choice experiment?
When utility is linear in attributes and in the cost attribute, the marginal willingness to pay for an attribute equals the negative ratio of that attribute's estimated coefficient to the cost coefficient, −β_a/β_cost. Intuitively, the cost coefficient measures the disutility of one currency unit, so dividing an attribute's preference weight by it converts the weight into money. Welfare measures for whole scenarios are computed by aggregating these attribute-level values.
What is the independence of irrelevant alternatives (IIA) and why does it matter?
IIA is a property of the basic conditional/multinomial logit: the relative odds of choosing between two alternatives do not depend on what other alternatives are present. It follows from the independent, identically distributed error assumption and can be unrealistic when alternatives are close substitutes. When IIA fails, ratios and welfare estimates can be distorted, so analysts use nested logit, mixed (random-parameters) logit, or latent-class models that relax it and also capture preference heterogeneity.
When should a choice experiment be preferred over contingent valuation?
Prefer a choice experiment when the good is naturally described by several attributes and you need the value of each attribute and the trade-offs between them, or when policy can vary attributes independently and you want to value scenarios not yet observed. Contingent valuation is often simpler and more direct for a single, holistic provision change — particularly when the dominant concern is a pure non-use (existence) value rather than attribute-level marginal values.
Sources
- 1.McFadden, D. (1974). Conditional logit analysis of qualitative choice behavior. In P. Zarembka (Ed.), Frontiers in Econometrics (pp. 105–142). New York: Academic Press.ISBN 9780127761503
- 2.Louviere, J. J., Hensher, D. A., & Swait, J. D. (2000). Stated Choice Methods: Analysis and Applications. Cambridge University Press.ISBN 9780521788304
You have read it. What now?
Cite this page
ScholarGate. (2026, June 22). Choice Experiment Valuation. ScholarGate. https://scholargate.app/economics/choice-experiment-valuation