Unskilled and Unaware: Subpar Users of Automated Driving Systems Make Spurious Decisions

Abstract

This work investigated differences between preference and performance in Human Computer Interactions and their dependency to the respective user skill level. A driving simulator study with N=57 participants was conducted to evaluate a Human-Machine Interface for a Level 3 Automated Driving System. Two experimenters rated interaction performance (e.g., input errors, mode confusions, etc.). Additionally, participants reported their preference by means of perceived usability and acceptance. The sample was split into four groups based on the quartiles of performance. Results revealed that the four groups differed significantly in their performance. However, preference ratings did not show this effect. Thus, the present research could find evidence that the dissociation of performance and preference depends on participants' skills. Finally, future research directions are outlined.