Effects on welfare measures of alternative means of accounting for preference heterogeneity in recreational demand models

Abstract

Multiattribute-revealed preference data are used to investigate heterogeneity in a sample of kayakers for a panel of whitewater sites in Ireland. This article focuses on a comparison of preference heterogeneity using a random parameter logit model with correlated tastes and a latent class model, in terms of the implications for welfare measures of environmental quality and site-access changes. Recreationalists' skill levels are found to affect preferences in both approaches. Statistics for the estimated distribution of welfare changes for the average respondent are computed for changes in site attributes, but contrary to previous work, these are found to be of similar magnitude.