USING FUNCTIONAL MEASUREMENT TO IDENTIFY THE FORM OF UTILITY FUNCTIONS IN TRAVEL DEMAND MODELS

Existing statistical procedures for model estimation that use data on observed behavior focus principally on the problem of estimating the parameters of the model, given the functional form. In contrast, methods for measuring function, which are psychological measurement procedures that use laboratory or interview data, provide a powerful set of tools for diagnosing the functional form of behavioral relationships. This paper explores the potential of a synthesis of these approaches in which functional measurement is used to guide travel demand model specification. A case study on choice of residential location by rural workers provides evidence that a model form based on functional measurement gives a better specification than the typical alternative functional forms used in travel demand models. Although it is relatively limited in scope, the case study strongly suggests that functional measurement methods can improve demand model specification. /Authors/