Leveraging mobile eye-trackers to capture joint visual attention in co-located collaborative learning groups

This paper describes a promising methodology for studying co-located groups: mobile eye-trackers. We provide a comprehensive description of our data collection and analysis processes so that other researchers can take advantage of this cutting-edge technology. Data were collected in a controlled experiment where 27 student dyads (N=54) interacted with a Tangible User Interface. They first had to define some design principles for optimizing a warehouse layout by analyzing a set of Contrasting Cases, and build a small-scale layout based on those principles. The contributions of this paper are that: 1) we replicated prior research showing that levels of Joint Visual Attention (JVA) are correlated with collaboration quality across all groups; 2) we then qualitatively analyzed two dyads with high levels of JVA and show that it can hide a free-rider effect (Salomon and Globerson 1989); 3) in conducting this analysis, we additionally developed a new visualization (augmented cross-recurrence graphs) that allows researchers to distinguish between high JVA groups that have balanced and unbalanced levels of participations; 4) finally, we generalized this effect to the entire sample and found a significant negative correlation between dyads' learning gains and unbalanced levels of participation (as computed from the eye-tracking data). We conclude by discussing implications for automatically analyzing students' interactions using dual eye-trackers.