Do Eye Movements Contribute to Accurate Metacognitive Judgments During Multimedia Learning?

• In Event: 35.031 - Multimodal Data During Learning With Advanced Learning Technologies: What Does Evidence Reveal About Self-Regulated Learning?

Abstract

Successful self-regulated learning (SRL) with multimedia (i.e., text and diagrams) requires students to accurately monitor the relevancy of the materials they are learning from (Azevedo, 2014; Mayer, 2014). Specifically, accurate content evaluations (CEs; judgments where students assess the relevancy of the materials to their current goal, Greene & Azevedo, 2009) are key for effective learning as they can direct students to learn more efficiently. In this study, we converge students’ eye-tracking data with their metacognitive judgments to investigate the relationship between eye movements and accurate metacognitive monitoring judgments. We focus on how the proportion of time students spend fixating on specific components of multimedia content (i.e., text, diagram, and science question) can be differentially associated with the accuracy of their CEs for text and diagrams.

45 undergraduates interacted with MetaTutorIVH (an intelligent virtual human-, trial-, multimedia-based learning environment) during 18 counter-balanced, randomized, self-paced trials that consisted of science questions, metacognitive judgment prompts, and multiple-choice questions (see Figure 1).

Students’ text and diagram content evaluations (CEs) were automatically collected and coded for accuracy by MetaTutorIVH, while their eye movements were collected by the SMI RED 250, with a sampling rate of 60Hz. Areas of interest (AOIs) were superimposed on the text, diagram, and science question. We calculated the proportions of time students fixated within each of these individual AOIs to use as our independent variables.

We used multi-level modeling for our analyses. Two fully unconditional models were conducted on students’ text and diagram CEs, which revealed there was sufficient within-subject variability for students’ text (σ2 = 0.09, z = 19.30, p < .001) and diagram CEs (σ2 = 0.18, z = 19.30, p < .001).

Results indicated that there were no significant associations between the proportions of time spent viewing the text, diagram, and science questions with students’ text CEs. However, results revealed a significant association between the proportion of time spent viewing the text and the proportion of time spent viewing the science question, but not the proportion of time spent viewing the diagram, with diagram CEs. Specifically, the longer proportion of time students spent fixating on the science question, the higher their diagram CEs. Additionally, the shorter proportion of time students spent fixating on the text contributed to lower diagram CEs.

Research investigating metacognitive monitoring processes during multimedia learning has primarily neglected to examine how eye tracking can be used to infer how students monitor their learning with multimedia. Our results demonstrate that students who spend more time attending to the science question they need to answer and less time attending to the corresponding text have more accurate diagram CEs. As such, our results have important implications for the use of multimodal, multichannel process data to investigate cognitive and metacognitive SRL processes during learning.

Authors

• James Lester, North Carolina State University

• Nicholas Vincent Mudrick, North Carolina State University

• Roger Azevedo, University of Central Florida