On Incorporating Affective Support to an Intelligent Tutoring System: an Empirical Study

Previous research studies have reported strong evidence that the emotional state of students may have a considerable impact on their learning. In this paper, we present an empirical study that evidences that it is possible to influence the user's affective state in a controlled way, by adapting the system's response. As part of this paper, we have analyzed the affective impact of varying the level of help provided in an existing Intelligent Tutoring System. Results show that it is possible to use classification approaches to predict positive and negative variations in dominance, valence, arousal, and performance to a reasonable level of accuracy.