Abstract

Recent advances in eye tracking technology have shifted the emphasis away from traditional static images or scenes to the study of dynamic stimuli. Although the aim is to maximise ecological validity, the use of such dynamic stimuli is not without challenges. The labour required in coding regions of interests (ROIs) frame by frame tends to limit researchers from using such stimuli. Moreover, it is not always clear what the relevant ROIs are when group comparisons are made. Here we present a first evaluation of a new data-driven method with the potential to address these issues. We measured political attitudes using questionnaires and used this to differentiate gaze behaviour between individuals with left and right-wing political orientations. A total of 45 participants watched 80 video clips of two left and two right wing politicians (20 per category). We will present data on how often the two political groups (left and right wing viewers) differ in their gaze position and what parts of videos clips are associated with such differences.