Investigating Determinants in the Acceptance of Automated Taxis: Evidence from Online Screen-based and Virtual Reality-based Stated Choice Experiments

Abstract

The introduction of Automated Taxis (ATs) has the potential to bring economic, environmental, and social benefits to the current transport system. The success of ATs, as a nascent technology, crucially depends on whether potential users adopt and accept to use them and what determinants will affect their acceptance of ATs. Despite relative novelty of the topic, there is already a relatively vast literature on the factors influencing the choice of owning and/or using general Autonomous Vehicles (AVs), but only few papers have dealt with Automated Taxis (ATs). Due to the absence of an AT real market to obtain Revealed Preference (RP) data, Stated Choice (SC) methods are typically an appropriate way to elicit travellers’ preferences. However, even though customised and careful designed, a major problem implicit in the SC experiments is the ‘lack of realism’, leading to the well-known hypothetical bias. This problem is particularly relevant and predominant when SC experiments are used to study innovative products, because, as amply discussed in the context of electric vehicles, respondents lack of experience with the product and often-insufficient knowledge has a significant impact in the preferences elicited with SC experiments. In line with this discussion, the aim of this PhD research consists in providing empirical, theoretical and methodological evidence to contribute to the state of the art to understand, measure and quantify the determinants in the acceptance of ATs and it proposes a methodology to implement SC experiments in immersive virtual reality (VR) environments. Four datasets are collected using the same instrument but applied online and embedded on the VR. Online data were collected in the UK and China, allowing for a comparison of the impacts across nations. Other two sets of data were collected among respondents in Newcastle using the online survey for one set of data and the VR for another set of data. A comparison between these datasets allows testing the impact of VR on the preferences elicited. Finally the impact of living in a city where AT systems do exist was also explored using the data collected in China. Many interesting results were found. For example UK respondents are willing to pay on average 5 times and 10 times as much as the WTPs of Chinese respondents to save one hour of travel time and are willing to pay on average between 2-11 times more than Chinese participants to save one hour of waiting time. On the other hand, no significant differences were found in the preference for in-vehicle features. Overall, results showed that the Chinese AT market is less elastic than the UK market to changes in level of services characteristics, in-vehicle futures and social conformity measures, with the exception of some population segments. Interestingly, the British AT market share is still significantly affected by the latent psychological construct Hedonic Motivations. As for the VR impact, perhaps the most interesting result, however, is that the attribute to measure AT adoption rate, which is a very problematic attribute in the online SC experiments (as proved in all researches conducted for other innovations), was significant when measured in the VR environment. Results need to be confirmed by further evidence, but are promising that VR might help achieve better results in particular to measure social conformity effects. The PhD research provides an extensive discussion of the willingness to pay measure for all the attributes tested and a comparison among the four datasets collected as well as with the values reported in the international literature. It provides also methodological guidelines on how to build SC experiments embedded into immersive VR settings and how to empirically carry out the experiments. Methodology-related, policy-related implications, and limitations of the current study are also discussed.