Description:
<jats:p> The potential of Intelligent Tutoring Systems (ITS) to influence learning may be greatly enhanced by the system’s ability to accurately assess a student’s cognitive state in real-time. For this to happen, interactions with, and reactions to, training content must be collected and assessed; data is then used to inform instructional adaptation within the system. Validated sensors are available for this purpose and have been shown to correlate with cognitive and affective states linked to learning. However, sensors used to inform student models are often expensive and impractical for wide-range use. In this paper the authors present a study evaluating the efficacy of using Emotiv’s Electroencephalogram (EEG) Affectiv Suite outputs to inform an ITS student model. In this experiment, seventy-three participants interacted with the Cultural Meeting Trainer (CMT), a web-based cultural negotiation trainer, while Emotiv’s engagement, short-term excitement, and long-term excitement metrics were indexed and logged. Our analysis assesses the quality of Emotiv metrics across one well-defined and two ill-defined scenarios. Results show consistent outputs across tasks and support further examination into the Emotiv’s ability to accurately track cognitive state in a learning environment. </jats:p>