The Reliability of Mode Choice Model Based on Stated Preference Data

The paper is to study the reliability of mode choice models based on stated preference data which were collected prior to the opening of a new rail station in Hiroshima. The external validity of the stated preference models was examined using the revealed preference data collected subsequent to the opening, so that thestated preference models were proved empirically to overestimate the share of the new alternative. Therefore, following three methods for the purpose of updating the stated preference models were employed; 1) adjusting the parameters of stated preference model by using the revealed preference data obtained together with the stated preference data, 2) combining the parameters of revealed preferenceand stated preference models (Bayesian updating), and 3) adjusting the stated preference data with the aid of transfer price data. Of all the approaches, the third updating method gave us the best results.