A Critical Review of Interdisciplinary Methods Used to Examine the Role of Emotions and Computer-Based Learning Environments

• In Event: 70.020 - Methods and Methodologies for Researching Emotions in Teaching, Learning, and Motivation

Abstract

Emotions play a critical role in learners’ ability to monitor and regulate their learning about challenging topics and domains while using computer-based learning environments (CBLE) (e.g., intelligent tutoring systems) (Azevedo et al, 2013; D’Mello et al., 2014; Graesser et al., in press; Lester et al, 2013). Emerging evidence indicates that emotions play a critical role in learning, performance, problem solving, and self-regulation with CBLEs, however, capturing emotions during learning with CBLEs has been predominantly studied using self-report measures. As such, the objective for researchers is to design CBLEs able to detect, model, and adapt to changes in learners’ emotional fluctuations in order to dampen negative emotions and facilitate positive emotions related to effective learning and performance. In order to accomplish this objective, CBLEs must: (1) detect and monitor learners’ emotions (e.g., confusion) in real-time (e.g., Calvo et al., in press), (2) embody theoretically- based rules (e.g., Pekrun, et al., 2011; Scherer, 2009) to identify and classify fluctuations in learners’ affective expressions, and (3) use pedagogical rules to provide contextually-relevant adaptive scaffolding and feedback (e.g., through artificial pedagogical agents) depending on the detected emotions and their potential impact on learning and performance.

In this presentation we will synthesize key findings from several studies using MetaTutor, an intelligent hypermedia learning environment for science learning. A typical MetaTutor study involves dozens of college students who participated in a two-day experiment (e.g., Azevedo et al., 2012; Azevedo & Chauncey Strain, 2011; Mudrick et al., 2014; Taub et al., 2014) to learn about the human circulatory system. They were instructed to use several SRL processes during their learning session (e.g., setting-relevant learning goals, using effective learning strategies). Participants were randomly assigned either to an adaptive or non-adaptive condition. The effectiveness of each condition is typically assessed based on analyses of participants’ 2-hour session with MetaTutor where we collected the following from each participant: self-report measures and several trace methods including eye-tracking (e.g., attention allocation and integration of multiple representations of information), video recordings of the face (for affect detection and classification), log-files (e.g., quiz results, summaries and metacognitive judgments, learner-agent dialogue), and physiological data (e.g., GSR).

Authors

• Roger Azevedo, North Carolina State University

• Michelle Taub, North Carolina State University

• Nicholas Mudrick, McGill University