Abstract:
We investigate the effect of respondents' heterogeneous preferences regarding the good to be valued in binary-choice contingent valuationstudies. This is especially important when the valuation options give rise to both highly negative and highly positive attitudes.The survey design employs questionnaires which include separate items for willingness to favorandwillingness to contribute in monetary terms to the provision of the good. In light of our analysis, eliciting respondents' differentpreferences for the project first, without monetary considerations, wouldimprove the WTP estimates as welfare measures. Our results alsosuggest that nonparametric estimation could offer another solution foraccounting for preference heterogeneity. Nonparametric estimation is less sensitive to those `no' responses which are in fact not intended toreflect the respondents' WTP as such but merely to express their dislikefor the good at issue. Copyright Kluwer Academic Publishers 2000