How, and for whom, will activity patterns be modified by self-driving cars? Expectations from the state of Georgia. (April 2020)

• Record Type: Journal Article
• Authors: Kim, Sung Hoo; Mokhtarian, Patricia L.; Circella, Giovanni
• Is Part Of: Transportation research. Volume 70(2020)
• Page Start: 68
• Page End: 80
• Publication Date: 2020-04
• Subjects: Autonomous vehicles -- Behavioral changes -- K-means cluster analysis -- Non-mean-centered factor scores -- Segmentation; Automobile drivers -- Psychology -- Periodicals; Automobile driving -- Psychological aspects -- Periodicals; Transportation -- Psychological aspects -- Periodicals; 629.283019
• DOI: 10.1016/j.trf.2020.02.012
• Languages: English
• ISSNs: 1369-8478

Abstract

Highlights: 16 AV-based activity changes yield 4 dimensions, e.g. distance, frequency, time use . Employed non-mean-centered (NMC) factor scores, preserving face meaning of neutral. Confirmed substantial discrepancy between traditional mean-centered and NMC scores. Six segments identified, e.g. change unlikely, longer trips, and more travel Leisure/long distance : wealthier, Atlanta. No change, change unlikely : older, rural. Abstract: Many studies have explored travelers' perceptions of self-driving cars (or autonomous vehicles, AVs) and their potential impacts. However, medium-term modifications in activity patterns (such as increasing trip frequencies and changing destinations) have been less explored. Using 2017–2018 survey data collected in the US state of Georgia, this paper (1) measures (at a general level) how people expect their activity patterns to change in a hypothetical all-AV era; (2) identifies population segments having similar profiles of expected changes; and (3) further profiles each segment on the basis of attitudinal, sociodemographic, and geographic characteristics. In the survey, respondents were asked to express their expectations regarding 16 potential activity modifications induced by AVs. We first conducted an exploratory factor analysis (EFA) to reduce the dimensionality of the activity-change vector characterizing each individual, and estimated non-mean-centered (NMC) factor scores (which have been rarely used in applied psychology). The EFA solution identified four dimensions of activity change: distance, time flexibility, frequency, and long distance/leisure . Next, we clustered Georgians with respect to these four-dimensional expectation vectors. The cluster solution uncovered six segments: no change, change unlikely, more leisure/long distance, longer trips, more travel, and time flexibility & more leisure/long distance . Using NMC factor scores identified considerably more inertia with respect to expectations for change than would have been apparent from the usual mean-centered scores. Finally, the various segments exhibit distinctive demographics and general attitudes. For example, those in the more leisure/long distance cluster tend to be higher income and are more likely to be Atlanta-region residents compared to other clusters, while those in the no change and change unlikely clusters tend to be older and are more likely to be rural residents.