A Method for Improving Welfare Estimates from Multiple-Referendum Surveys

This paper introduces an alternative non-market value elicitation method—the “quasi-doublereferendum†(QDR)—applied to barrier island restoration in Mississippi. It is appropriate for surveys that elicit willingness-to-pay responses to multiple projects differing in scale only and can be used to increase efficiency while mitigating bias. We compare results to the more commonly used single-referendum (SR) method under two admissible ranges of willingness to pay: negative to positive infinity, and zero to income. The confidence intervals of the QDR models were narrower. We argue that the QDR approach should be less subject to bias than the commonly used double-referendum approach.