Treatment of “Don't-Know” Responses in Contingent Valuation Surveys: A Random Valuation Model

Abstract This paper provides a utility-theoretic interpretation of “Don't-Know” (DK) responses in contingent valuation surveys using a referendum elicitation procedure. The random valuation model in this paper assumes that each respondent of a contingent valuation survey has an implicit valuation distribution, rather than a single true value as traditionally assumed, in his/her mind. If an offered referendum bid is not clearly different from the mean value of one's valuation distribution, the respondent may give a DK answer to the CV referendum question. Based on this model, a maximum likelihood procedure is proposed for estimating respondents' willingness to pay for a nonmarket good with referendum contingent valuation data which include DK responses. An empirical illustration of this estimation procedure is also provided.