Reading focus
Syllabus: How Do We Judge and Decide?
Follow how the same alternatives can receive different rankings when the response procedure changes. Identify the assumptions behind treating an elicited value as a stable measurement.
Why you're reading it. Establish preference construction as a question about process and measurement before considering particular choice models.
A question to consider. If choosing and pricing two travel options produce different rankings, what would you investigate before declaring one response the traveler's true preference?
Why this reading
Asking someone to choose an option and asking how much that option is worth can produce different rankings. Slovic uses this problem to examine what a preference is and how researchers can measure it. The argument makes the act of asking part of the phenomenon that needs explaining.
Before you read
A preference ranks alternatives. Elicitation is the procedure used to obtain that ranking, such as choosing, pricing, or matching one option to another. Procedure invariance means that equivalent elicitation methods preserve a preference order. Utility represents the value an option has for a chooser; it is not simply its monetary amount.
Read for
- Reconstruct a preference reversal and identify what changes between the two elicitation procedures.
- Compare explanations involving compatibility, prominence, and reasons for choosing. What evidence would distinguish them from random response error?
- Examine the move from laboratory findings to preference construction. Ask when stable, familiar preferences might remain and what would count against a broad constructive account.
Consider afterwards
- A student chooses a quieter room but assigns a higher price to a larger room. Design a comparison that could reveal whether response format changes the weight given to each feature.
Editorial context
This award-address synthesis argues for a constructive view, particularly for complex or unfamiliar choices. It does not establish that people have no enduring values or that every preference can be changed at will. Its discussion of practical applications also raises a responsibility for researchers: a measurement procedure may shape the response it purports to record. That implication requires transparent elicitation, not an assumption that one method uncovers a uniquely true preference.