Abstract
A model system that predicts travel mode and party size jointly, is developed. This model system accounts for the correlation between party size and mode choice that arises due to omitted variables that affect the two. The coefficients of mode choice utility functions are assumed to vary with party size. The parameters are estimated using activity diary data and applying the bivariate probit model. The parameter estimates indicate that mode choice behavior is associated with party size, party size is affected by person attributes and the trip purpose, and the value of time increases with party size.