Understanding Attitudes Towards Self-Driving Vehicles: Quantitative Analysis of Qualitative Data

Self-driving vehicles represent potentially transformative technology. But achieving this potential depends on people’s attitudes towards this technology and willingness to use it. Ratings from surveys estimate acceptance, and open-ended comments provide an opportunity to understand the “why” behind the ratings. One way to understand the content of open-ended comments is through computer-based text analytics. A recent survey of 8,571 nationally representative drivers in J.D. Power’s 2017 U.S. Tech Choice StudySM included a rating of willingness to use self-driving vehicles and an associated open-ended response. We present a quantitative analysis of these qualitative, open-ended responses that uses structural topic modeling to reveal the basis of the respondents’ attitudes. Drivers’ attitude towards self-driving vehicles was quite negative: only 11% stated they “definitely would” trust self-driving technology, whereas 35% stated they “definitely would not.” The structural topic modeling identified 10-topics, such as “Many unknowns” and “Don’t trust” that help explain these negative attitudes.