An exploration of heterogeneity in Latent Psychological Variables on Travelers' destination choice

Discrete choice models have been widely used to model individuals' decision-making processes in numerous contexts. Given that people's tastes and preferences over alternative shopping destinations vary for reasons not necessarily observed by demographic characteristics, more intricate modeling techniques are required to capture taste variations. Latent factors, such as the psychological impact on the destination choice process, can explain heterogeneity among individuals. This study considers the latent class difference in shopping destination choices by using empirical data collected from 812 individuals for two types of grocery and clothing in Tehran, Iran. The heterogeneity in destination choice behavior is captured by a Mixed Logit (ML) model and a Latent Class Variable (LCV) model formulation. A simple multinomial logit model has also been developed to benchmark the two techniques. Our findings suggest that both the ML and the LCV can capture taste variation in travelers' behavior, where the ML expectedly provides a better fit to the data.