Evidence suggests the calibration of hypothetical and actual behavior is good-specific. We examine whether clustering commodities into mutual categories can reduce the burden. While we reject a common calibration across sets of commodities, a sport-specific calibration function cannot be rejected.

This paper calibrates real and hypothetical willingness-to-accept estimates elicited for consumer goods in a multi-unit, random nth-price auction. Using a within-subject experimental design, we find that people understated their real willingness to accept in the hypothetical regimes, framed both...