Bridging the digital divide in urban mobility: a latent class choice analysis of older adults’ preferences for autonomous vehicles in China

As China experiences rapid population aging, ensuring inclusive and sustainable mobility for older adults has become a pressing policy concern. Autonomous vehicles (AVs) offer the potential to enhance mobility independence, yet their adoption among older adults remains uneven due to heterogeneous psychological dispositions and digital capability constraints, particularly in app-mediated AV systems. This study examines how digital capabilities and behavioral mechanisms jointly shape older adults’ preferences for private and shared AVs, using a hybrid modeling framework. A mixed-mode survey of 996 older adults in Wuhan, China was conducted, combining a stated preference experiment, Theory of Planned Behavior (TPB) indicators, and a five-domain digital capability scale aligned with DigComp. A sequential hybrid approach was applied: latent psychological constructs were first estimated using Structural Equation Modeling and subsequently incorporated into a Latent Class Choice Model (LCCM) to capture unobserved preference heterogeneity. The results identify two distinct segments. A public-transport-preferring group (54.82%) exhibits strong cost and time sensitivity, limited AV experience, and lower procedural digital skills, while an AV-preferring group (45.18%) shows higher baseline propensity toward AVs, greater willingness to pay for time savings, and stronger capabilities in communication and content creation. Behavioral intention and perceived behavioral control are robust predictors of AV choice, whereas subjective norms exert a consistently negative influence and trust in automation does not emerge as a stable driver once usability and capability constraints are considered. Elasticity and willingness-to-pay estimates further reveal that price-based interventions are most effective for the transit-preferring segment, whereas time and convenience improvements drive responses among AV-preferring users. These findings highlight the importance of segment-specific AV policies, emphasizing simplified access and targeted pricing for digitally constrained users, and service-quality and usability improvements for digitally proficient older adults.