Choice Behavior Analysis for Feeder Modes on First and Last Mile with Qualitative Factors

Feeder modes on First and Last Mile recently have been getting attention. To carry out relevant policies, it is necessary to understand the choice behavior about feeder modes on first and last mile. Therefore we analyzed traveler’s mode choice behavior about electric scooter, autonomous bus, sharing bicycle and walking which can be introduced on First and Last Mile. On the mode choice behavior, the qualitative factors have influence as well as the quantitative factors. Therefore, this study quantified the qualitative factors which were implied “feeling danger” related to bicycle and electric scooter, and “concerning health” related to walking. The values of qualitative factor were estimated by using a binary logit model with socio-economic attributes and personal preference as independent variables, and the mode choice models were calibrated in the form of a multinomial logit model with traveler’s socio-economic attributes, mode service attributes, and qualitative factor as independent variables. The influence of the qualitative factors on traveler’s choice behavior was interpreted as unit of travel time. And the calibration analysis of transportation mode choice behavior at very short distances showed that the model is statistically significant. The calibrated model explained that the people in age 20s preferred to electric scooters comparing with other age groups and that houseperson tended to avoid walking. The passengers tended to feel danger to electric scooter more than bicycle. The effect of “feeling danger” and “concerning health” on the mode choice behavior in time unit, their effect was found as same as between 3.17 and 5.06 minutes. When considering that it is a very short-distance moving, the results showed that the qualitative factors of “feeling danger” and “concerning health” affect greatly on the traveler’s mode choice behavior. The study contributed to quantifying the qualitative factors for traveler’s mode choice behavior at very short distances. This made it possible to express the influence of qualitative factors numerically.