A cross-nested logit model for trip type-mode choice: An application

We applied the cross-nested logit (CNL) model to home-based trips reported in the 2001 National Household Travel Survey (NHTS) database to explain individuals' choices of trip type and mode to use. We also compared the CNL model results to those from the nested-logit (NL) model and the network Generalized Extreme Value (Network GEV) model. Alternative modes (private, public, walk and other) are available to each of the four trip purposes (home-based work, home-based shopping, home-based recreational, and homebased other), namely the nests. The results indicated that the CNL model conveniently models decisions on trip type and mode to use made by individual travelers. Its results are comparable to those of the network GEV model. The models identify the impact of individuals' socioeconomics and alternative attributes on trip type and mode choices. The presence of children in a household increases the probability of choosing a private mode while it decreases the probability of using public transport or walk. Private mode and public transport were more attractive to travelers than walking when considering the combined trip distance and travel time. Individuals from households with higher income prefer private mode to public transportation. However, income is not a significant factor on a decision to walk. Since the data used is a national-wide database, a study in which a different data set collected locally is recommended. The NHTS data set may suffer from spatial variation among the travelers and hence produce inconclusive results.