Modelling the Effect of Travel Experiences in Modal Choice Using Virtual Reality and Physiological Sensor Data

Abstract

The effect of experiences on travel mode choices is well established in the literature. Additionally, there is evidence that psychophysiological signals, such as skin conductance, can capture travel experiences without relying on self-reported measures, given their strong correlation with psychological states. However, using physiological data to estimate the effect of experiences on choices remains unexplored due to challenges in data collection. The advent of virtual reality (VR) presents a unique opportunity to gather such data under controlled laboratory conditions and explore how travel experiences shape future demand. This paper presents three waves of VR experiments where participants repeatedly chose between different travel modes, including current (car, bus, ride-hailing) and futuristic options (autonomous vehicle, air-taxi, hyperloop). After making their choice, they experienced the mode in the VR environment, and indicated whether they would have preferred another option. This is the first experiment to analyse psychological states and modal choice within a VR environment, and the first to use physiological data to assess how experienced psychological states affect future choices. We estimate a dynamic hybrid model that accounts for the effects of lagged latent stress and inertia. Galvanic skin conductance is used to measure latent stress. Our findings show that driving in VR was the most stress-inducing option, reducing the likelihood of repeating that choice. Additional results, methodological implications, and the potential of VR for modal choice studies are discussed.

Keywords: virtual reality, physiological data, latent stress, travel experience, dynamic hybrid model