Modifying willingness to pay estimates where respondents mis-report their preferences

The likelihood for the Logit model is modified, so as to take account of uncertainty associated with mis-reporting in stated preference experiments estimating willingness to pay (WTP). Monte Carlo results demonstrate the bias imparted to estimates where there is mis-reporting. The approach is applied to a data set examining consumer preferences for food produced employing a nonpesticide technology. Our modified approach leads to WTP that are substantially downwardly revised.