Eye tracking and early detection of confusion in digital learning environments: Proof of concept

Abstract

Research on incidence of and changes in confusion during complex learning and problem-solving calls for advanced methods of confusion detection in digital learning environments (DLEs). In this study we attempt to address this issue by investigating the use of multiple measures, including psychophysiological indicators and self-ratings, to detect confusion in DLEs. Participants were subjected to two intrinsically confusing insight problems in the form of visual digital puzzles. They were asked to solve problems while their eye trajectories were recorded and these data were triangulated with self-ratings of confusion and cued retrospective verbal reports. All participants had a significant increase in fixations on relevant (i.e., related to the solution) and not-relevant areas at an early stage of the problem-solving process. However, only fixations on not-relevant areas were positively correlated with confusion ratings. Moreover, participants who significantly solved the problem differed in their fixations duration on relevant and not-relevant areas from non-solvers. The importance of early detection of confusion and the affordances of emerging technologies for this purpose are discussed.