Comparison of Mode Choice Models Calibrated with Network Data and Reported Data

This paper presented the effects of travel data and choice set selection on the estimation results of mode choice models. The paper seems to suggest that reported travel data would give better results than network travel data. Besides these findings, reported travel data have some other advantages. First, network travel data are determined in part by the coding practices while reported travel data are free from it. Second, survey method, coding practice, and limitation of the network travel data are generally unknown to the analysts. Third, in an area with an out-of-date network travel data or without any travel data, collecting reported travel data may be more cost effective for some special purposes. The paper also showed that the selection of choice set had a great effect on the model results.