Some Issues in Continuous- and Discrete-Response Contingent Valuation Studies

I want to begin with a caveat. I have no training in psychology, and my interest in the field is that of an amateur. Thus, the confidence with which I shall make statements about psychological facts is in inverse proportion to my expertise. My point of departure is the assumption that the most serious problem with contingent valuation (CV) methodology is hypothetical bias. Apart from introspection, the evidence for this assumption comes from two classic studies conducted by Richard Bishop and his colleagues at the University of Wisconsin. The first study by Bishop and Heberlein (1979) involved a survey of two random samples of individuals who had obtained permits to hunt ducks in the Horizon Zone of East Central Wisconsin. One sample group was asked the hypothetical question whether they would be willing to sell their permit for a specified sum of money (this sum differed among individuals). The other sample group was given a real opportunity to sell their permit for a specified sum of money; of 237 hunters surveyed, 105 actually sold their permits.