DYNAMIC ANALYSIS OF STATED PREFERENCE, FOR THE NEW TRANSIT SYSTEM OF LIGHT RAIL TYPE DUE TO BE OPENED IN 1994 IN HIROSHIMA, WAS CARRIED OUT USING THE PANEL DATA COLLECTED FOR WORK/SCHOOL TRIPS AT TWO POINTS IN TIME; 1987 AND 90. IT WAS FIRST EMPIRICALLY DEMONSTRATED THAT THE STATED PREFERENCE FOR THE NTS IS GREATLY AFFECTED BY CURRENT TRABEL MODES CONSISTING OF CAR AND BUS, AND THAT THE TEMPORAL CHANGES OF STATED PREFERENCE WHICH WERE CAUSED BY THE CHANGE OF CURRENT TRAVEL MODE DO NOT SUPPORT THE MARKOVIAN ASSUMPTIONS WHICH INDICATE THE INDIVIDUAL'S RESPONSES TO A CHANGE OF TRAVEL ENVIRONMENT OVER THE THREE YEARS. THE IMPORTANCE OF STATE DEPENDENCE WAS ALSO DEMONSTRATED BY BUILDING DYNAMIC SP MODE CHOICE MODELS OF MULTINOMIAL LOGIT TYPE.
