Modeling and analysis of freight mode choice behavior integrating grouped and repeated observations

Modeling and understanding freight mode choice behavior is vital for policy makers and operators to design environment-friendly freight systems and improve service quality. This paper studies the shippers’ choice between truck and road-rail intermodal. Considering that a single shipper provides multiple groups of stated preference data under different transport contexts, a model integrating grouped and repeated observations (IGRO) is constructed based on a standard mixed logit (ML) model to capture the heterogeneity of intra-individual. A standard multinomial logit (MNL) model, a cross-sectional ML model and a standard panel ML model are also tested in this study. Statistical indicators show that the IGRO model performs the best, which indicates that the intra-individual heterogeneity across different transport contexts is stronger than the correlation. The potential factors contributing to the heterogeneity of intra-individual freight value of time (FVOT) are different origin-destinations (ODs) and cargo categories. In addition, a two-stage weighted calibration scheme is proposed. Subsequently, based on the calibrated model, some suggestions for increasing the intermodal market share are obtained through elasticity calculation and scenario analysis. Moreover, this paper measures carbon emission reduction under different modal shift scenarios and its monetary values.