This paper addresses issues more directly related to the external validity of stated preference (SP) data. It reports some early results from a study, which investigated the temporal stability and inter-temporal properties of SP data; this study used longitudinal data from a before-and-after investigation of the effects of introducing the Trondheim Toll Ring in Norway. Some of the key issues and opportunities, arising in the analysis of longitudinal SP data, are discussed, and existing evidence is reviewed about the temporal stability and inter-temporal properties of such data. The background of the Trondheim study is described; its SP exercise involved drivers' responses to twelve toll ring timing timing and pricing scenarios, each of which was described in terms of four characteristics. The main results of the Trondheim analysis of temporal stability and inter-temporal properties are presented, including results from the basic model for all car drivers and for commuters, and from the extended model for commuters. Although the results suggest that the temporal stability of SP-based tastes and preferences cannot be assumed, temporal variation in SP-based tastes seems to be related to changes in socio-economic and travel circumstances.
Abstract