Trip End Mode Split

Trip interchange models and their underlying assumption that relative level of service (LOS) of travel modes is a major determinant of mode choice for the work trip are examined. The investigation is done through a comparative analysis of trip end and trip interchange mode split for work trips in the greater Toronto area (GTA). Two model sets are developed; one includes trip end models and the other includes trip interchange models. The models are developed using a relatively new statistical procedure for nonparametric analysis of data referred to as classification and regression trees. The results show that the explanatory power enjoyed by trip end models is virtually the same as that of trip interchange models. Relative LOS is found to have an insignificant role in explaining heterogeneity in mode split, particularly outside the central business district of the study area. It is recommended that predictions made by trip interchange models of mode shift following changes in LOS attributes be approached with caution. Trip end models, which attribute less significance to relative LOS than trip interchange models and are thus conservative with respect to mode shift, could be considered in planning studies, particularly in areas of low transit use. The interaction between the decisions of residential and employment locations, car ownership, and mode choice is emphasized, and the need for more research to investigate and model such interactions is stressed.