摘要：Recent attempts to test the validity of the contingent valuation method have relied on laboratory-type experiments. In these experiments, willingness to pay responses in hypothetical choice experiments are compared with responses from choice experiments requiring actual payments. Often evidence of hypothetical bias is found. Critical for these experimental tests of hypothetical surveys is that the methodology used to elicit willingness to pay from subjects in the real-payment experiment be demand revealing. If it is not, then differences in responses to hypothetical and real valuation questions could be due to free-riding in the real-payment survey and not due to hypothetical bias in the hypothetical survey. This paper reports on experiments that implement a theoretically incentive-compatible revelation mechanism (a closed referendum) to elicit responses to valuation questions in both hypothetical and real experiments. As in earlier studies, evidence of an upward hypothetical bias is found.
