Incentive Compatibility in an Attribute-Based Referendum Model

The manner in which WTP survey responses are elicited has received much attention in the nonmarket valuation literature because of the potential bias that may be introduced via alternate response formats. One issue of particular concern is that of incentive compatibility. Several studies have concluded that response formats that present a series of valuation questions are not incentive compatible. While the single dichotomous choice elicitation format reduces the amount of information collected, multiple-bounded elicitation formats may yield biased, unreliable results. Understanding this trade-off can help provide better information on response formats that will elicit incentive compatible responses and therefore provide realistic and policy-relevant information. This research investigates the effect of the number of multinomial choices presented to respondents in an attribute-based referendum (ABR) format on incentive compatibility of responses. Data was collected from two versions of a mail survey that used an attribute-based description of a hypothetical forest easement program in Michigan's Upper Peninsula and a referendum-style choice between the status quo and various forest easement program scenarios. The first version of the survey presented four choice scenarios to respondents while the second version presented only one choice scenario. Results suggest that the multiple-bound response format improves statistical efficiency due to the statistical significance of all estimated parameters. However, this efficiency may also overestimate WTP. The single-bounded response format displays lower statistical efficiency but may reflect more accurate preferences from respondents. Results lead to a rejection of the hypothesis that the number of choices presented to respondents has no effect on results and have implications for the reliability of nonmarket valuation information from multiple-bound response formats in attribute-based referenda models