How Should Consumers' Willingness to Pay Be Measured? An Empirical Comparison of State-of-The-Art Approaches

TLDR

The study compares four common willingness‑to‑pay measurement methods—open‑ended, choice‑based conjoint, Becker‑DeGroot‑Marschak, and incentive‑aligned conjoint—to real purchase data. Using a five‑in‑one design, the authors benchmark each method against real purchases, evaluating statistical fit and decision‑relevant performance. BDM and incentive‑aligned conjoint satisfy statistical and decision tests, reveal higher price sensitivity and many none‑choices in incentive‑aligned settings, and suggest that open‑ended and standard conjoint can still yield accurate demand curves despite hypothetical bias.

Abstract

This study compares the performance of four commonly used approaches to measure consumers’ willingness to pay (WTP) to real purchase data (REAL): the open-ended question format (OE), choice-based conjoint analysis (CBC), the incentive-compatible mechanism proposed by Becker, DeGroot, and Marschak (1964) (BDM), and incentive-aligned choice-based conjoint analysis (ICBC). With this five-in-one approach, we test the relative strengths of the four measurement methods, using REAL as the benchmark, on the basis of statistical criteria as well as decision-relevant metrics. Our results indicate that BDM mechanism and incentive-aligned conjoint analysis can pass statistical and decision-oriented tests. We find respondents to be more price sensitive in incentive-aligned settings than in non-incentive-aligned settings and the real purchase setting. We further find a large number of none-choices under incentive-aligned conjoint analysis in comparison to hypothetical conjoint analysis. Our study also uncovers an intriguing possibility: even when the open-ended question format and choice-based conjoint analysis generate hypothetical bias, they may still lead us to the right demand curves and right pricing decisions.