Sequences of Frustration and Confusion, and Learning

Abstract

In this paper, we use sensor-free affect detection [4] and a discovery with models approach to explore the relationship between affect occurring over varying durations and learning outcomes among students using Cognitive Tutor Algebra. Researchers have suggested that the affective state of confusion can have positive effects on learning as long as students are able to resolve their confusion [10, 22], and recent research seems to accord with this hypothesis [17]. However, there is some room for concern that some of this earlier work may have conflated frustration and confusion. We replicate these analyses using sensor-free automated detectors trained to distinguish these two affective states. Our analyses suggest that the effect may be stronger for frustration than confusion, but is strongest when these two affective states are taken together. Implications for these findings, including the role of confusion and frustration in online learning, are discussed.