Abstract
The tourism-driven traffic has generated a huge amount of carbon dioxide emissions and raised the attention of environmental sustainability. This study investigates tourists' preferences on mode choices at a green destination in Taiwan. We consider both transportation- and tourism-related attributes to examine preferences among five alternatives. A stated-preference questionnaire is developed to collect 383 valid responses in terms of the principle of systematic sampling. Regarding tourists' heterogeneous behaviors, Mixed Logit (ML) and Latent Class Model (LCM) are utilized to analyze the collected responses. The results show that all adopted attributes are statistically significant. More importantly, LCM can exceptionally outperform conventional multinomial logit model (MNL) and ML. LCM successfully segments tourists into three groups and reveals distinct effects from marriage and income status of tourists. Then corresponding parameters of individual groups are estimated and a simulation further reveals the conversion among modes when different transportation policies are implemented.
