Exploring Effects of Using Different Forms of Travel Time in Mode Choice Behaviour Modelling for Peak Hours in Workdays: A Case Study of xi'an, China

Travel time has always been an important factor in modelling travel mode choice behaviour. Different forms of travel time can be used in the modelling, such as the exact travel time of each mode, travel time difference (TTD) between two modes, and the ratio of one mode’s travel time to the benchmark mode. Different forms of travel time may have different effects in mode choice behaviour modelling. Using 2015 household travel survey data from Xi’an in China, This study developed multinomial logit (MNL) models and nested logit (NL) models to investigate the effects of different forms of travel time in mode choice behaviour modelling for peak hours in workdays. The results indicated that time difference is the most important time indictors in mode choice decision. Therefore, it is necessary to incorporate TTD into the evaluation criteria to assess the quality of public transit service. Even though the result was derived from the data in Xi’an, it can also be useful for other fast-developing cities in policy-making to encourage the use of public transport thus to make the entire transport system more sustainable.