Abstract
It is apparent that travel behavior is different across individuals. If homogeneity of the population assumed in con-ventional analysis is not accepted, observed heterogeneity should be considered in the model. If individual omitted variables (e. g. attitude and taste) influence the parameter estimation of model, unobserved heterogeneity becomes a serious problem in model building. In this paper, we classify the individual heterogeneity into observed and unobserved ones and check the exsistence of observed heterogeneity for mode choice models using market segmentation me-thod. Besides, we examine the effectiveness of nonparametric Mass Point approach which can introduce unobserved heterogeneity and which does not depend on the distribution of heterogeneity.
