Should all choices count? Using the cut-offs approach to edit responses in a choice experiment

Should we give equal weight to all responses in a choice experiment? Previous Choice Modelling papers have considered the issue of the internal consistency of choices-such as the extent to which strictly-dominated options are chosen. In this paper, a different focus is employed, namely the extent to which people choose options which violate their stated upper or lower limits for the acceptable levels for individual attributes. Since hypothetical over-statement of WTP has been a focus in stated preference studies, we concentrate on violations of stated upper limits of WTP, and explore the effects of "editing" such choices using a variable censoring rule and alternative approaches to re-classifying choices. The empirical case study is a choice experiment on eco-tourists in Rwanda. Our main conclusion is that the suggested approach offers a useful way of imposing consistency on choices, and that editing choices in this manner has an appreciable impact on estimated willingness to pay. However, issues remain with regard to explaining why people apparently violate their stated maximum willingness to pay.