Abstract
This report provides a preliminary validation of the demand model used in the TRRL programme of research into dial-a-bus. It compares the observed ridership of the experimental Harlow dial-a-bus system with predictions obtained from a modal split model. It is shown that the predicted patronage was about one-third too high but that the predicted distributions of the age of the passengers, their journey purposes and the modes foregone, were in reasonable agreement with the observations. These results indicate that the basic structure of the demand model is sound, but that, in the modelling, too low a generalised cost had been assumed for dial-a-bus. (Author/publisher)