This paper investigates alternative methods to account for preference heterogeneity in choice experiments. Heterogeneity can be explained by assuming its influence to impact the systematic component of utility, the stochastic one or both. Seven different models were estimated to search for heterogeneity along the three dimensions described. The comparison based on model performance and willingness to pay measures shows that methods to search for heterogeneity are not independent one from the other and might produce substantially different results giving rise to different policy implications.
