TRANSPORTATION MODE CHOICE:

ARE LATENT FACTORS IMPORTANT?

Abstract

Many past researches have invoked time, cost and specific mode attributes to explain transportation mode choice. Alternatively, this paper attempts to study an additional dimension by incorporating latent factors. The general structure comprises two models, namely a latent variable model and a discrete choice model. The estimation is performed by first determining the significance of specific latent variables— safety, comfort, convenience and flexibility. In the second stage, the discrete choice model is estimated by integrating the choice and the latent variable in the choice process. The discrete choice model is estimated via logit and probit regressions. The main findings are that the latent factors are statistically significant and thus enrich the discrete choice model. Hence, latent factors must not be ignored in describing individual preferences towards travel mode. By providing better explanation to mode choice behaviour, the enriched discrete choice model could prove useful to policy makers in managing transportation infrastructure and services to maximize welfare.