This paper presents a two-stage Stated Preference survey to investigate the impacts of travel time reliability and in-vehicle crowding on the mode choice decisions across four different transport modes, i.e. car, metro, park and ride (P&R) and bus. The decisive attributes considered are average travel time, travel time reliability, cost and in-vehicle crowding. Five model specifications are defined for the parameter estimations. Significant interaction effects between in-vehicle crowding and travel time are found. Time multipliers are defined to represent the effects of in-vehicle crowding. In contrast, no evidence could be established for the interaction between in-vehicle crowding and travel time reliability. Results of the mode-specific valuations of travel time reliability and in-vehicle crowding, vary remarkably across the four different transport modes. In the mode-specific models, the range of time multipliers is estimated to be [1.44, 2.00]. Besides, demand estimates would be biased when the mode-specific willingness to pay (WTP) is ignored. For instance, the mode share of metro will be underestimated when its reliability level is high, and vice versa. This suggests that mode-specific WTP of travel time reliability and in-vehicle crowding should be considered in the demand estimations and in the earlier stage of public transport project appraisal.