Modal choice models estimation using mixed Revealed and Stated Preferences data

Discrete choice demand models estimation can be made using information on travel behavior obtained by Revealed and Stated Preferences surveys. Revealed Preferences techniques, traditionally utilized, are relative to the actual users travel behavior in a real context. Stated Preferences techniques are methodologies based on statements made by interviewees about their preferences in different choice contexts, real, hypothetical or experimental. Therefore, an important innovation is introduced: the possibility to consider choice alternatives not available at the time of the surveys. These techniques have been adopted for demand models calibration to predict the choices made by users and their preferences variations while the choice context changes. Particularly, some Multinomial Logit mode choice models have been specified and calibrated. The calibrated models have been distinguished in: RP models, based on the choices made by users exclusively in the real context; SP models, based on the choices stated in the hypothetical contexts and joint RP/SP models, using RP and SP data on the same sample. The study has confirmed the utility of SP techniques for users travel behavior analysis in a hypothetical choice contexts; moreover, it has confirmed, as attended, than RP and SP conjoint analysis improves parameters estimation in discrete choice models.