Do subjective measures of transit mode attributes better explain commute satisfaction than objective ones? Empirical evidence from commuters in the Seoul Metropolitan Area

The present study investigates determinants affecting commute satisfaction in the Seoul Metropolitan Area, South Korea. In particular, two types of mode attributes including objective (actual commute time and average crowdedness) and subjective (satisfaction with them) measures are incorporated to identify which indicators play a more meaningful role in determining commute satisfaction. Developing both conventional ordered probit models and more advanced models with random parameters, this research addresses two research questions related to the limitation of objective measures in terms of the model fitness and taste heterogeneity among commuters. The results show that subjective measures do not necessarily guarantee a better explanation power, meaning that either indicator can be a source of commuters' disutility. However, subjective measures appear to be a better indicator of subjective well-being in terms of interpretability, as we confirm that (1) parameter values vary after allowing for heterogeneity at the individual level and (2) individual coefficients of objective measures possibly have a mixed impact (both negative or positive) on the dependent variable. It suggests that subjective measures are worthwhile for modellers, practitioners and decision-makers to explore behaviour at a deeper level although it may not dramatically contribute to model improvement.