Design criteria to develop choice experiments to measure the WTP accurately.

To measure the willingness-to-pay (WTP) accurately, Vermeulen et al.[2008] apply the c-optimality criterion to generate designs for conjoint choice experiments. This criterion is based on minimizing the sum of the variances of the WTP estimators approximated by the delta method. Designs generated based on this criterion lead to more accurate WTP estimates than the ones obtained by standard designs and reduce considerably the occurrence of extreme WTP estimates, although they do not exclude them. In this paper, other optimality criteria are considered to tackle this problem. We distinguish between criteria in preference space on the one hand and criteria in WTP-space on the other hand. In a simulation study and a numerical example, we compare the accuracy of the WTP and the utility coefficient estimates yielded by the designs based on these new criteria.conjoint choice experiment; Bayesian optimal design; willingness-to-pay; conditional logit model;