Unraveling hypothetical bias in discrete choice experiments

Our study contributes to the literature on hypothetical bias in discrete choice experiments in two main respects. First, using stated and revealed preference data collected from a sample of grocery shoppers, we explore the effectiveness of two mitigation techniques- the Bayesian truth serum and the use of indirect questions (inferred valuation). Results show that both techniques are able to reduce (but not completely remove) hypothetical bias. Second, we contribute to the understanding of sources of hypothetical bias. We find evidence that signaling is a source of hypothetical bias in direct questions and affects public attributes. Our findings also suggest that biased perceptions of one's own and/or other people's values are a source of hypothetical bias in direct and indirect questions and affect public and private attributes. (C) 2020 Elsevier B.V. All rights reserved.