| Sumario: | A study was conducted using recently developed statistical techniques to derive enhanced parameter estimates from models of dichotomous choice contingent valuation data. Data were drawn from the study by Bateman et al. of respondents' willingness to pay a particular monetary value to prevent saline flooding in the Norfolk Broads, a unique freshwater wetland area situated near the North Sea in the East Anglian region of England. Random effects associated with bid amounts presented to respondents were modelled alongside the variance of individual responses to produce a nested model that more accurately reflects the structure of the data being studied. Quasi-likelihood methods for estimating parameters in such hierarchical models are discussed, and a simulation method for estimating goodness of fit is demonstrated. However, as estimated parameters may still be biased, a parameter bootstrap technique is presented and compared to a delta method approximation. The findings suggest that these methodological techniques may be employed together with theoretical advances in contingent valuation to inform the ongoing debates regarding choice of appropriate welfare measure and optimal study design.
|
|---|