Covariates-dependent random parameters regret–rejoice models of choice behavior: specification and performance assessment using experimental design data

Abstract

This paper elaborates the classic regret-minimizing model introduced in transportation research by adding rejoice. Regret is defined as a function of attribute differences between the considered and the best alternative, and as a function of attribute differences between the considered alternative and all other choice alternatives. Rejoice is defined as a function of attribute difference between the best and the second best choice alternative. To allow for heterogeneity in regret-rejoice, the parameters of the model are assumed to be a function of a set of socio-demographic covariates, plus pure random taste variation. Using a stated choice experiment of the choice of shopping center and a revealed mode choice data set extracted from the Dutch national travel survey, results indicate that the proposed regret-rejoice models outperform the regret-only models and that the inclusion of social demographic covariates in estimated parameters considerably improves the performance of all models.