A Link Between Trust in Technology and Glance Allocation in On-Road Driving

Abstract

Summary: This paper examines whether there is an association between preexposure trust in technology and subsequent glance behavior when interacting with a technology that was relatively novel for the majority of participants. After rating their level of trust in technology on a questionnaire, participants drove one of two vehicle models on a highway and engaged in a voice-based navigation address entry task. Subjective ratings of trust in new car technologies were found to be significantly positively correlated with a higher frequency of glances across all coded glance regions during the task. In one of the voice-interface implementations, these higher ratings of trust were also associated with a higher frequency of glances to the user interface, but with fewer long duration (>2s) glances per minute. A lower trust in technology in general showed some association with taking more time to complete interactions. The findings are discussed as highlighting the potential value of further research into the associations between trust and visual scanning behavior.

Trustworthiness Perceptions in Code Review

“…Previous work reports that participants' trust significantly impacts various eye-tracking metrics over the system [8,9,22] However, even though participants' self-report acknowledges a bias against both machines (readability) and humans (quality), we find no impact of the patch's provenance on participants' high-level performance. With regard to overall behavior, such as acceptance rates or time spent, participants were not influenced by provenance.…”

“…Much previous work has focused on selfreporting (i.e., subjective scales) such as think-aloud protocols, surveys, and interviews to measure trust [22], which suffer from the Hawthorn (observer) effect [4,5] and may not be reliable in a software maintenance context [7,15]. A few recent studies have used eye tracking as an objective, biologically-based measure to provide insights into the cognitive processes and the perception of trust in a continuous, non-subjective and non-intrusive manner [8,9,22]. These studies suggest a correlation between users' trust and their visual scanning behavior.…”

The influence of system transparency on trust: Evaluating interfaces in a highly automated vehicle

“…The current study demonstrated the presence of motion sickness because participants were staring at the Animation and Camera Feed interfaces. Future research could analyse in more detail where people's visual attention is (Geitner et al, 2017) and correlate this information with the resulting user discomfort (Smyth et al, 2020) to inform the design interfaces that attract user attention only when required.…”