Abstract
IN THIS ARTICLE THE AUTHORS CONSIDER WHETHER THE COEFFICIENTS OF TRAVEL MODE-CHOICE MODELS ARE TRANSFERABLE TO AREAS OTHER THAN THOSE FOR WHICH THEY WERE ESTIMATED. PREVIOUS STUDIES HAD INDICATED THAT MODEL BIAS WAS OFTEN DUE NOT ONLY TO DIFFERENT POPULATIONS, BUT TO DIFFERENCES IN SAMPLING, DATA COLLECTION PROCEDURES AND MEASUREMENT OF VARIABLES. THE EMPIRICAL STUDY CONTROLS FOR THESE SOURCES OF BIAS AND ANALYSES DATA SETS WERE BASED IN TWO SYDNEY SUBURBAN AREAS. THE RESULTS SUGGEST THAT MODEL COEFFICIENTS ARE NOT TRANSFERABLE.