Scenario Realism and Welfare Estimates in Choice Experiments
Evidence from a study on implementation of the European Water Framework Directive in Denmark

When using stated preference data the researcher makes specific assumptions about the respondents whereby respondents are assumed to understand and accept the policy scenario presented to them. It is further assumed that respondents are capable of evaluating different alternatives proposed to them and that they make utility-maximizing choices. We suggest that respondents may disagree with survey information provided. Not controlling for these dispersed beliefs concerning survey information could bias the welfare estimates in an unknown direction. We test for such biases using data collected to assess the potential welfare gains of implementing the European Union Water Framework Directive. The survey information is based upon scientific evidence in terms of monitoring data for the environmental status of a water body. Our findings confirm our hypothesis in that we cannot reject that welfare measures that do not control for dispersed beliefs are biased.