Modeling Travel Tool Ownership of the Elderly Population: Latent Segmentation-Based Logit Model

Abstract

A latent segmentation-based logit (LSL) modeling framework investigated the travel tool ownership of the elderly population in the greater Toronto area, Ontario, Canada. This study utilized data from the 2006 Transportation Tomorrow Survey conducted in the greater Toronto area. The study developed five mutually exclusive travel tool ownership combinations or bundles for older adults. One of the unique features of this study is the development of an LSL model to capture unobserved heterogeneity in the older adults’ travel tool ownership decisions. The LSL model captures unobserved heterogeneity by allocating individuals to discrete latent segments through a latent segment allocation model. The model results suggest that the LSL model is estimated for two latent segments in which less frequent trip makers are more likely to belong to Segment 1 and more frequent trip makers have a higher probability of belonging to Segment 2. Moreover, the model results suggest that considerable heterogeneity exists among the older adults in the two segments. For instance, older adults residing in a neighborhood with a higher percentage of owned dwellings prefer to own multiple travel tools in one latent segment and reveal an opposite relationship in another segment. Moreover, older adults living closer to regional business centers have a higher probability of owning a monthly transit pass in one segment and exhibit an opposite relationship in another segment. This diversity in travel behavior should be addressed within the transportation and land use policies to ensure an effective and equitable transportation system for older adults.