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

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