Assessing the Impact of Distance Modelling Approaches on Active Travel Predictions

Abstract

Trip distance is widely considered a key determinant of active travel choice, although distance is handled in choice modelling in various ways. This study examines how linear, categorical, quadratic, and piecewise representations for distance influence mode choice. The piecewise approach demonstrates superior fit and provides more stable elasticity estimates, ranging from -1% to -3%, compared to the broader range of -0.17% to -36% observed in categorical treatments and occasional positive values in quadratic treatments. These results underscore the piecewise model's advantage in capturing critical behavioural shifts and offering more reliable insights to inform policy.