Does respondent uncertainty explain framing effects in double bounded contingent valuation?

Many stated preference studies have reported framing effects in responses to valuation questions. This occurs when respondents use irrelevant information contained in a question to help them value the good. We investigate if respondent uncertainty can explain two commonly observed framing effects in contingent valuation studies. Specifically using a double bounded dichotomous choice elicitation format, we investigate anchoring (or starting-point bias) and the shift effect, which may indicate if the method is incentive compatible. Respondent uncertainty is measured using a follow up question that asks respondents their certainty about their valuation. We ?nd evidence that the anchoring effect is stronger for respondents expressing uncertainty about their valuation compared to respondents expressing certainty. The shift effect is significant and negative only for respondents expressing certainty. Our findings suggest that anchoring can be reduced if respondents are certain of their valuation, and that iterative elicitation formats are not incentive compatible