Captivity and Choice in Travel-Behavior Models

The correct calibration of individual-choice models and the correct prediction of choices from such models can only be achieved by restricting data sets to those individuals who have choices. This paper discusses the ideas of choice and captivity with respect to travel choises, and explores the effects on calibration and prediction of including captives in the data. It is shown that the inclusion of captives causes biases in the coefficients and predictions that cannot be removed a posteriori and that these biases have important policy implications for highway and transit planning. Hence, the restriction of data to choosers alone is important in both the development and the use of individual-choice models of travel behavior.