Exploring passenger assessments of bus service quality using Bayesian networks

Studies on public transit have emphasized the role of passenger satisfaction with service quality in travel choice decisions and indicated that satisfaction depends on various service attributes. Few studies have, however, systematically examined the underlying relationships among service attributes to assess their influence on passenger overall satisfaction. Therefore, to contribute to this rapidly-emerging literature, this paper applies Bayesian networks to quantify the influence of each service aspect on passenger overall satisfaction with regular bus service quality. This analysis involved 609 passengers who participated in a 2013 regular bus service survey in Nanjing, China. The derived Bayesian network shows the relationships among service attributes and passenger overall satisfaction graphically. In particular, service aspects such as running on schedule, acceptable waiting time, available seats, clean onboard environment, pleasant environment at stations, convenient design for transfers, and air-conditioning were the key determinants of overall satisfaction with bus service.