Online and in-lab final sample structure.

This study investigates sociotechnical imaginaries of autonomous vehicles (AVs) using a dual approach: in-lab and online eye-tracking experiments. We examine how cognitive engagement varies across hypothetical decision-making scenarios involving algorithmic failure of AVs. In comparison with non-AV scenarios. This article highlights the characteristics, advantages, and limitations of methods, emphasizing their complementary contributions to understanding how individuals perceive and engage with emerging technologies. The in-lab experiment revealed high-quality and precise data from a homogeneous sample, while the online experiment enabled us to scale the research and explore diverse sociotechnical imaginaries from a global sample through crowd-sourced platforms. Key findings show that both in-lab and online participants exhibited longer gaze durations at one point, predominantly longer in AV scenarios. However, a deeper analysis of overall cognitive engagement revealed that in-lab participants, with more concentrated sociotechnical imaginaries, were more focused on non-AV scenarios, indicating a stronger emphasis on human decision-making. In contrast, online participants, whose imaginaries may be shaped by global perspectives and diverse experiences with data and algorithms, displayed increased attention toward AV scenarios, with significant visual variations among participants, reflecting global interest or concern over high-stakes algorithmic decisions. These findings contribute to our understanding of how perception of AVs differs globally and offer insights into emerging concerns around algorithmic decision-making in everyday life.