The role that affective states play in learning was investigated from the perspective of a constructivist learning framework. We observed six different affect states (frustration, boredom, flow, confusion, eureka and neutral) that potentially occur during the process of learning introductory computer literacy with AutoTutor, an intelligent tutoring system with tutorial dialogue in natural language. Observational analyses revealed significant relationships between learning and the affective states of boredom, flow and confusion. The positive correlation between confusion and learning is consistent with a model that assumes that cognitive disequilibrium is one precursor to deep learning. The findings that learning correlates negatively with boredom and positively with flow are consistent with predictions from Csikszentmihalyi's analysis of flow experiences.

Two experiments explored the integration of animated agents into multimedia environments in the context of R. E. Mayer's (2001) cognitive theory of multimedia learning. Experiment 1 was a 3 (agent properties: agent only, agent with gesture, no agent) ϫ 3 (picture features: static picture, sudden onset, animation) design. Agent properties produced no significant effects. Both sudden onset and animation conditions facilitated performance relative to the static-picture condition. In Experiment 2, we explored the effects of printed text, spoken narration, and spoken narration with the printed text, in a multimedia environment that included an agent, to investigate effects of redundancy. The spoken-narration-only condition outperformed the other 2, with no differences between printed text and printed text with spoken narration.

We explored the reliability of detecting a learner's affect from conversational features extracted from interactions with AutoTutor, an intelligent tutoring system (ITS) that helps students learn by holding a conversation in natural language. Training data were collected in a learning session with AutoTutor, after which the affective states of the learner were rated by the learner, a peer, and two trained judges. Inter-rater reliability scores indicated that the classifications of the trained judges were more reliable than the novice judges. Seven data sets that temporally integrated the affective judgments with the dialogue features of each learner were constructed. The first four datasets corresponded to the judgments of the learner, a peer, and two trained judges, while the remaining three data sets combined judgments of two or more raters. Multiple regression analyses confirmed the hypothesis that dialogue features could significantly predict the affective states of boredom, confusion, flow, and frustration. Machine learning experiments indicated that standard classifiers were moderately successful in discriminating the affective states of boredom, confusion, flow, frustration, 123 46 S. K. D'Mello et al. and neutral, yielding a peak accuracy of 42% with neutral (chance = 20%) and 54% without neutral (chance = 25%). Individual detections of boredom, confusion, flow, and frustration, when contrasted with neutral affect, had maximum accuracies of 69, 68, 71, and 78%, respectively (chance = 50%). The classifiers that operated on the emotion judgments of the trained judges and combined models outperformed those based on judgments of the novices (i.e., the self and peer). Follow-up classification analyses that assessed the degree to which machine-generated affect labels correlated with affect judgments provided by humans revealed that human-machine agreement was on par with novice judges (self and peer) but quantitatively lower than trained judges. We discuss the prospects of extending AutoTutor into an affect-sensing ITS.

No abstract

No abstract