View article

This paper investigates the validity of the hypothesis that the appropriate specification of the hedonic function should use mean levels of environmental quality. A test of the mean specification is constructed by estimating a hedonic equation with the probabilities of various levels of environmental quality entered as independent variables. The test indicates rejection of the mean specification. The results imply that benefit estimates based on the traditional mean model are likely to be biased and that efforts to improve the accuracy of hedonic methods should consider more complete measures of environmental quality.