Combining discrete and continuous representations of preference heterogeneity: a latent class approach

This paper investigates heterogeneity in preferences for forest recreators in Mallorca, Spain. We develop a latent class approach combining discrete and continuous representations of tastes and compare it with the conventional latent class and random parameter logit approaches. We investigate the performance of the discrete-continuous model by comparing welfare estimates and predictive accuracy. The discrete-continuous model outperforms latent class and mixed logit approaches when comparing goodness-of-fit and in- sample site-choice forecasts. We find that the discrete-continuous model for preference heterogeneity reveals variation among individuals' preferences and WTP, and for some policy changes our results reveal striking differences in means and distributions of WTP.