A theoretical and statistical exploration into the effects of morals, personality and uncertainty on hypothetical bias in contingent valuation

Contingent valuation (CV) is a surveying technique used to estimate the willingness to pay (WTP) by individuals for non-market goods (goods for which no market of exchange exists). Many CV studies have shown the existence of hypothetical bias which occurs when individuals are asked a hypothetical WTP question followed by a request for actual donations, and the real payments are not the same as (and usually less than) the hypothetical WTP. This dissertation examined hypothetical bias using both theoretical and statistical analysis. The theoretical analysis consisted of a literature review and a microeconomic exploration of utility theory. The literature review identified past studies that expanded on Walrasian theory of consumer behavior including the ideas of option and existence value, fair share, and social approval. Other articles reviewed discussed uncertainty with respect to supply of a resource as well as the preferences and income of the consumer. The microeconomic theory section evaluated the concepts from previous literature within a utility framework that in turn was used to guide the statistical analysis. The statistical analysis examined two empirical case studies. Each study consisted of a set of three surveys administered to several college-level classes. The surveys collected information on the socioeconomic characteristics of respondents and asked for each respondent\u27s WTP (in a donation mechanism) for a specific good. The commitment level was tightened and respondents were asked for an actual donation. In the first case study respondents were asked for their WTP for a scholarship fund and the second case study respondents were asked for a donation to a fund purchasing clean air permits. Results of the case studies found hypothetical bias of about 2650%. When the question wording changed to include more information and a higher level of commitment, the mean WTP was not statistically different at the 5% level. Only one respondent between either study actually donated money. The Mach IV test, when used as a personality proxy, may be an important factor in estimating WTP and hypothetical bias. The mean WTP between different levels of certainty was not statistically different at the 5% significance level. Finally, many students indicated that they might be making their decisions without considering economic trade offs and that respondents\u27 MP may also be influenced by some aspects of personality, social approval and other characteristics not associated with the good being valued