Multi-Attribute Utility Theory for Preference Construction: Impacts on Consistency Across Stated Preferences Methods

Stated preference techniques are often required to incorporate valuations of nonmarket goods into welfare analysis. Some authors defend a constructive approach to preferences elicitation, claiming that individuals form their valuations during the elicitation process rather than relying on pre-existing, well-defined and stable preferences. This perspective has motivated proposals to use a Multi-Attribute Utility Theory (MAUT) framework to guide respondents through a structured process of preference construction. This paper test the virtues of a MAUT information treatment to improve internal consistency between two stated preference methods-the Adaptive Bisectional Dichotomous Choice (ABDC) and Payment Cards (PC)-which have been found to yield different results in this context. In an online survey applied to a representative sample of the adult French population, one group of respondents was assisted to reflect on trade-offs between income, health, and social interactions before completing the ABDC and PC tasks. We find that the treatment increased the likelihood of observing monotonic preferences and reduced the incidence of zero willingness-to-pay (WTP) responses. However, it did not reduce the gap between methods; if anything, the discrepancy increased, partly due to the lower share of zero WTP responses in the treatment group. Besides, the treatment imposed a higher cognitive burden. Overall, the MAUT-based information treatment changed the pattern of stated preferences in ways consistent with more deliberate preference construction, but it did not remove their dependence on the elicitation method.