Destination choice set definition in travel behaviour modelling

Abstract This paper addresses conceptual and practical problems with the use of probabilistic models that insist on generalized choice sets for destination choice modelling. It is proposed that classification of destinations (functional destinations) be used to define a common choice set of types of destination choice behaviour. For shopping travel, these functional destinations are classifications with respect to shopping destination attraction measures and their spatial separation (in terms of distance, travel time or accessibility) from trip origins. Crossed, nested and nested crossed functional destinations are discussed. A multinomial response relation model is presented as a convenient means of predicting patterns of functional destination classifications. Empirical analyses on crossed functional destinations are discussed. The need for market segmentation with respect to some attitudinal factors is established in terms of model parameter plausibility and statistical testing. A procedure for modelling destination choice behaviour, starting from functional destinations down to the assignment of specific destinations, is presented.