What experimental protocol influence disparities between actual and hypothetical stated values?

Abstract

Preferences elicited in hypothetical settings have recently come under scrutiny, causing estimates from the contingent valuation method to be challenged due to perceived "hypothetical bias." Given that the received literature derives value estimates using heterogeneous experimental techniques, understanding the effects of important design parameters on the magnitude of hypothetical bias is invaluable. In this paper, we address this issue statistically by using a meta-analysis to examine data from 29 experimental studies. Our empirical findings suggest that on average subjects overstate their preferences by a factor of about 3 in hypothetical settings, and that the degree of over-revelation is influenced by the distinction between willingness-to-pay and willingness-to-accept, public versus private goods, and several elicitation methods.