ELICITATION AND TRUNCATION EFFECTS IN CONTINGENT VALUATION STUDIES

The contingent valuation method (CVM) uses surveys of expressed prefer ences to evaluate willingness to pay for (generally) non-market, envir onmental goods. This approach gives the method theoretical applicabili ty to an extensive range of use and passive-use values associated with such goods. However, recent years have seen the method come under sus tained empirical and theoretical attack by critics who claim that the expressed preference statements given by respondents to CVM questions are subject to a variety of biases to the extent that ''true'' valuati ons cannot be inferred. This debate was reviewed and assessed in the r ecent report of the US, NOAA ''blue-ribbon'' panel which gave cautious approval to the method subject to adherence to a rigorous testing pro tocol. This paper reports findings from the first UK CVM study to gene rally conform to those guidelines. The major objective of the research reported on here is the analysis of the effects of altering the metho d of eliciting willingness to pay (WTP) responses. Three WTP elicitati on methods are employed: open-ended questions (where the respondent is free to give any answer); dichotomous choice questions (requiring a y es/no response regarding a set WTP bid level); and iterative bidding q uestions (where a respondent is free to move up or down from a given W TP starting point). Results indicate that respondents experience signi ficant uncertainty in answering open-ended questions and may exhibit f ree-riding or strategic overbidding tendencies (although this is less certain). When answering dichotomous choice questions respondents seem to experience much less uncertainty although the suggestion that bid levels affect responses cannot be ruled out, and it is clear that resp ondents behave somewhat differently to dichotomous choice as opposed t o open-ended formats. The iterative bidding approach appears to provid e a halfway house with respondents exhibiting certain of the character istics of both the other formats. We concluded that the level of uncer tainty induced by open-ended formats is a major concern, and that furt her research into the microeconomic motivations of individuals respond ing to iterative bidding and dichotomous choice CV surveys is high pri ority. A further aim of the analysis was to test for changes in estima ted mean WTP induced by the application of different forms of truncati on across all elicitation methods. Recommendations are made on appropr iate truncation strategies for each elicitation method.