Monitoring Affective Trajectories During Complex Learning

Abstract

Monitoring Affective Trajectories during Complex Learning Sidney D’Mello (sdmello@memphis.edu) Department of Computer Science, University of Memphis Memphis, TN 38152 USA Roger S. Taylor (rstaylor@memphis.edu) Institute for Intelligent Systems, University of Memphis Memphis, TN 38152 USA Art Graesser (a-graesser@memphis.edu) Department of Psychology, University of Memphis Memphis, TN 38152 USA A series of studies have recently explored the affective states that occur during complex learning. Studies by Graesser and his colleagues have collected online measures of affect, such as observations by trained judges and emote- aloud protocols, as well as offline judgments of emotions by multiple judges (Craig, et al., 2004; D’Mello et al., 2006; Graesser et al., 2006). These studies have revealed that the basic emotions identified by Ekman and Friesen (1978), namely anger, fear, sadness, joy, disgust, and surprise, typically do not play a significant role in learning (see also Kort Reilly, & Picard, 2001). Instead they documented a set of affective states that typically do play a significant role in learning, at least in the case of college students learning about computer literacy with an intelligent tutoring system. These affective states were boredom, flow (engagement, Csikszentmihalyi, 1990), confusion, and frustration. They also monitored the affective states of delight and surprise, which occurred less frequently. While some of these affective states might be viewed as purely cognitive in nature, our position is that they should be classified as affective states (or emotions) because these states are accompanied by significant changes in physiological arousal compared with a “neutral” state of no apparent emotion or feeling (Barrett, 2006; Meyer & Turner, in press; Stein & Hernandez, in press). Furthermore, affective-cognitive composites are particularly relevant to higher-order learning. The aforementioned set of affective states can be situated within a broader perspective of emotion, in particular Russell’s (2003) Core Affect framework. This perspective holds that an affective state is composed of two integrated components: valence (pleasure to displeasure) and arousal (activation to deactivation). These components can be depicted graphically with valence represented on the X-axis and arousal on the Y-axis. Moving from left to right along the X-axis (valence) would correspond to increasing feelings of pleasure. Moving upward along the Y-axis (arousal) would correspond to increasing feelings of activation and energy (see Figure 1). The affective states of boredom, flow, confusion, and frustration will be the primary focus of this paper. These affective states have been previously correlated with learning Abstract This study investigated the transitions between affective states (i.e., boredom, flow, confusion, frustration, delight, and surprise) during learning while college students were tutored in computer literacy by AutoTutor, an automated tutoring system with natural language dialogue. Videos of participants’ faces and the interaction histories were recorded and then played back for the participants to judge their own affective states. We developed a metric to measure the relative likelihood of transitioning from an affective state at time t i to a subsequent affective state at time t i+1 . Several significant trajectories between affective states were identified. Instructional implications are discussed in the context of an expanded version of a cognitive disequilibrium model. Keywords: Affective states; emotions; affect trajectories, affect sequencing; emotion dynamics; AutoTutor; learning; instruction. Introduction There is ample empirical evidence in the psychological literature that emotions (or affective states) are systematically influenced by the knowledge and goals of the learner, and vice versa (Mandler, 1984; Ortony, Clore, & Collins, 1988; Russell, 2003; Stein & Levine, 1991). As individuals interact in the social and physical world, they attempt to assimilate new information with existing knowledge schemas. When new or discrepant information is detected, a mismatch between input and knowledge occurs. Attention shifts to discrepant information, the autonomic nervous system increases in arousal, and the new information may modify the individual’s goals and knowledge. The learner experiences a variety of possible emotions, depending on the context, the amount of change, and whether important goals are blocked. However, this type of affective arousal that accompanies learning is still not well understood. For example, researchers have yet to narrow down the emotions that accompany deep level learning of conceptual material. The consequential impact of the emotions on knowledge acquisition and transfer is still not well understood.