This empirical study presents an analysis of mode choice for selected urban trips in the San Francisco Bay area. The economic model is a restricted consumer choice model, where the mutually exclusive collectively exhaustive choice is between auto driver and transit passenger. The main testable hypothesis is that, in the absence of knowledge about the value a traveler attaches to his time, if a choice exists and if a mode is cheaper than the alternative in terms of both time and money, it should be chosen. The hypothesis was subjected to empirical analysis and, within the limits of the data, appeared to be a good approximation to reality. The statistical model used to further investigate the data was discrimination-classification analysis. The discriminant function can be interpreted as an indifference hypersurface of the indirect or constrained utility function. The statistical theory underlying the three versions of the model used is presented along with a derivation of the elasticity of choice. The data were a subset of the Bay Area Transportation Study Commission origin and destination home interview data merged with interzonal travel times and costs of both modes. Trips were stratified by purpose. Elasticities were calculated and compared both within a purpose by different variables and between purposes for each variable. Some potential policy changes were treated in the context of the model and compared with results of other studies.