Abstract:Contingent valuation methodology (CVM) has been increasingly applied to cultural resources. CVM employs survey methods to gather stated preference information, which can be used to estimate economic values of various cultural resources and projects. Although popular in other fields, the application of CVM in the cultural arena is relatively recent. This article summarizes this growing body of empirical literature and its range of findings. A meta-analysis gives a statistical view of the "state of the art" of the literature. This preliminary analysis sheds light on the consistency and validity of the use of this method in cultural applications. First, CVM is briefly and formally described. Second, a history and survey of the literature captures the breadth of research published thus far and highlights some findings. A few representative studies are described in greater detail. Third, the broader universe of CVM studies applied to cultural resources is described and analyzed in quantitative terms using a preliminary meta-analysis. 1 This meta-analytical approach allows for a more systematic accounting of the nature of these studies. Finally, the conclusion summarizes where the "state of the art" of CVM stands in the field of cultural economics.
Keywords
BackgroundCVM involves using surveys to elicit a willingness to pay (WTP) from individuals for hypothetical changes in some good or service. CVM is the most popular method in a family of alternative stated-preference techniques, known as "Choice Modeling" or "Conjoint Analysis."Generally, these surveys ask respondents to rate, rank, or choose among alternatives described in terms of various levels of attributes (Mazzanti, 2002). This paper considers all of these closely related methods but refers to "CVM" for brevity.The basic theoretical framework for CVM studies considers an individual (or household) utility function, u, as a function of market goods, x, and the level of a public good or service provided, q. where Q is a vector of survey characteristics, β and δ are vectors of regression coefficients, and ε is the error term. Statistical significance of β tests survey validity for variables that economic theory predicts would determine demand (income, substitutes, etc.). Likewise, tests of significance of δ can indicate sensitivity to changes in scope of q or other survey biases. The regression allows the estimation of a mean WTP, calculated at the sample or population mean of the independent variables collected from the survey.
3The dichotomous choice format, like a referendum, asks ("yes" or "no") whether WTP exceeds a specified amount. Econometrically, the probability of a "yes" response for a given dollar amount w (called a "bid" amount) is modeled as Pr(yes) = Pr(WTP + ε > w). The WTP response is typically regressed on a constant, w, X, and Q using logit, probit, or Weibull functions (see Giraud et al., 2001). This allows the calculation of a mean WTP parametrically at sample means.