Joint modeling of mode choice and travel distance with intra-household interactions

Travel mode and distance choices are not independent decisions, and individual choices are usually made in the knowledge of the preference and needs of other household members. To represent the dependency between mode choice and travel distance and the role of intra-household interactions in travel behavior, we develop a joint discrete-continuous model with intra-household interactions. In this model, we use joint household tour as the unit of analysis and characterize the dependency between mode choice and travel distance using flexible copula functions. Then, we apply the proposed model to analyze tour-based mode choice and travel distance in Beijing, China. The results indicate that unobserved factors contribute to positive dependency between mode choice and travel distance. The choice of walk is more dependent on travel distance than other travel modes, and the choice of travel modes except public transit shows a higher correlation with travel distance in complex individual tours than in simple individual and joint household tours. Further, a comparison between the proposed model, the independent model, and the model without intra-household interactions reveals that ignoring the dependency between mode choice and travel distance, or not considering intra-household interactions, could lead to over- or under-estimation of the effects of changes in exogenous variables.

Automated summary

Research has found that the choices of travel mode and distance are not independent decisions, as individuals usually consider the preferences and needs of other household members. The dependency between mode choice and travel distance is influenced by unobserved factors. Walk mode choice is more dependent on travel distance compared to other travel modes, and the correlation between mode choice and travel distance is higher in complex individual tours. Ignoring this dependency or not considering intra-household interactions may lead to over- or under-estimation of the effects of changes in exogenous variables.