Who is more likely to buy electric vehicles?

To promote electric vehicles, it is vital to know what impacts the preferences for electric vehicles over conventional fuel-based cars. To address this, a discrete choice experiment is developed and integrated into a survey. An online survey was conducted in Canada with 2062 valid responses. Different labels are designed for the survey to determine the most effective GHG information framing to increase the influence of such information on decisions. In this study, the influence of lifecycle emissions is considered. Three ensemble learning techniques are applied and they are compared based on prediction accuracy, and the most accurate technique is applied to determine the relative influence of variables on the intention to buy electric vehicles. Further, the interaction of variables is investigated using xgbfir. Subsequently, Accumulated Local Effect (ALE) is employed to examine the influence direction of top variables on the electric vehicle purchase likelihood. The results suggest that environmental attitudes and purchase price are the most influential parameters on the intention to buy electric vehicles. Moreover, those who are extremely worried about climate change, do not own a car, and self-identified as being at the top of the climate change stage of change are more likely to buy electric vehicles.

Uncontrolled Keywords

electric vehicle; ensemble learning; variables analysis; accumulated local effects