Using respondent uncertainty to mitigate hypothetical bias in a stated choice experiment

In a choice experiment study, willingness to pay for a public good estimated from hypothetical choices was three times as large as willingness to pay estimated from choices requiring actual payment. This hypothetical bias was related to the stated level of certainty of respondents. We develop protocols to measure respondent certainty in the context of a choice experiment, and to calibrate hypothetical choices using these certainty measures. While both the measurement of respondent certainty and the use of certainty measures to calibrate responses are complicated by the multiple-choice nature of choice experiments, calibration successfully mitigated hypothetical bias in this application.