Investigating the determinants of travel mode choice across age classes in Langsa, Indonesia utilizing logit model

Abstract

There are different preferences in the decision-making process of humans due to stochasticity. Therefore, this study was conducted to investigate the preferences in selecting a particular mode of travel. This involved using discrete choice modeling. The predictive performance of the model was also evaluated with the contribution of each variable to the model. This is useful for stakeholders to evaluate which factors have significant contributions enabling them to adjust policy accordingly. This study made use of surveys which incorporate revealed and stated preferences in the City of Langsa, Aceh, Indonesia to produce 13 variables including trip attributes and socio-demographic characteristics. This study employs tree distinguished models based on age classes within the sample: all-data, old-age class, and young-age class. Seven variables namely trip frequency, willingness to travel frequency, level of education, household transport expenditure, number of family members, travel cost , and travel time exhibit significancy in every model albeit with diverse extents. With negative vectors, travel cost appears to have the greatest magnitude of scale parameter among variables in every model. Furthermore, each model managed to predict the outcome of alternative 1 extremely well, scoring nearly a perfect 100% a piece. However, no model yields a good accuracy rate in predicting alternative 2, with all models scoring below 15%. All models exhibit good overall accuracy rates, correctly predicting in at least 7 out of 10 times.