Incentive-Compatible Preference Elicitation
Tie stated choices to real consequences to reveal genuine preferences
- Difficulty
- Easy
- Time to result
- ~days to results
- Steps
- 6
- Confidence
- 98%
Incentive-compatible elicitation makes expressed preferences consequential. Instead of merely asking what someone would choose, the researcher explains that one of the participant's decisions will be selected and implemented for real money or another valued outcome. Because any answer may determine what the participant receives, misrepresenting a preference creates a risk of being stuck with an unwanted result. The technique does not guarantee honesty, but it aligns truthful reporting with self-interest and helps expose hypothetical bias. A clean implementation defines the choices, establishes a credible selection procedure, supplies the promised budget or reward, and completes the selected transaction. Comparing these consequential decisions with hypothetical responses can also reveal where surveys systematically overpredict voting, purchasing, or risk-taking behavior.
Origin
Camerer presented this as a core commandment of experimental economics and illustrated it with paid risk assessments and real product purchases. Extracted from Masters in Business.
Core principles
- 01Stated intentions become more credible when consequences are real
- 02Reward people according to their expressed choices
- 03Select a decision for implementation so every answer may matter
- 04Design incentives around something participants genuinely value
How to run it
- 1
Specify consequential choices
Define the decisions participants will make and the real outcomes associated with each option.
Pro tip Use choices participants understand and genuinely care about.
Watch out Symbolic or trivial rewards may not change reporting behavior.
- 2
Fund the outcome
Provide money, products, or another valued resource sufficient to implement the selected choice.
Pro tip Keep stakes meaningful without making them coercive.
- 3
Explain the selection rule
Tell participants that one of their decisions will be chosen and carried out after the task.
Pro tip Use a simple random procedure that participants can trust.
Watch out Ambiguous implementation rules weaken incentive compatibility.
- 4
Collect independent decisions
Record each choice as though it could become the consequential one.
Pro tip Avoid feedback that lets later answers strategically offset earlier ones.
- 5
Implement the selected choice
Select one response using the declared rule and deliver its financial or material consequence.
Pro tip Document fulfillment so the procedure remains credible in future sessions.
Watch out Failing to honor the outcome invalidates the incentive structure.
- 6
Measure hypothetical bias
Where useful, compare consequential choices with answers from an otherwise similar hypothetical condition.
Pro tip Keep wording and options consistent across conditions.
Watch out Do not attribute every difference to dishonesty; attention and comprehension may also change.
In the wild
Participants first report whether they would hypothetically buy products at stated prices. In a consequential condition, they receive money and learn that one selected yes response will require an actual purchase, with the remaining money and product delivered to them.
→ Researchers can distinguish casual survey approval from willingness to complete a real transaction.
Common mistakes
Using rewards people do not value
An outcome cannot motivate truthful disclosure if participants are indifferent to it.
Making consequences unbelievable
Participants will continue answering hypothetically if they doubt that any choice will actually be implemented.
Confusing incentives with certainty
Consequences reduce hypothetical bias but do not remove misunderstanding, noise, or every form of strategic behavior.
Is it for you?
Best for
Experiments and product tests where at least one stated choice can be implemented for real.
Not ideal for
High-stakes settings where implementing a randomly selected choice would be unsafe, unethical, or unaffordable.
From the transcript
“we almost always pay people money or something we know they value based on the decisions they made.”
“we're going to pick one of the things you said you wanted”
“if you don't tell us what you really wanted, you're gonna get stuck with something.”
From the episode
Advancing Behavioral Economics with Colin Camerer
Colin Camerer