How Can Accurate Metacognitive Judgments Predict Successful Multimedia Learning?

• In Event: Advances in Understanding Metacomprehension Accuracy of Expository Text

Abstract

Objectives: Accurate metacognitive monitoring is required during multimedia learning to achieve higher learning outcomes (author, 2014). Limited research examining metacognitive monitoring during multimedia learning has found few metacognitive judgments to be predictive of successful learning outcomes (e.g., overconfident JOLs; Serra & Dunlosky, 2010). We argue that examining other metacognitive judgments (i.e., content evaluations [CEs]; author & colleague, 2009, retrospective confidence judgments [RCJs], Nelson & Narens, 1990) can inform us of monitoring processes that are more indicative of successful learning with multimedia. For this study, we focus on different metacognitive judgments than those previously examined to identify how accurate metacognitive judgments can predict increased learning outcomes during multimedia learning.

Theory: Multimedia learning theories suggest learners cognitively process information from text and diagrams separately and in different ways (Mayer, 2014). Evidence also suggests learners exhibit a bias toward text-based information during multimedia learning (Eitel & Scheiter, 2015). As the cognitive processes involved in selecting, organizing, and integrating information are different for text and diagrams, it should be expected that metacognitive judgments for these representations will also be different.

Methods and Results: 48 college students’ metacognitive judgments (i.e., CEs, RCJs) and multiple-choice responses were collected as they participated with the MetaTutor Learning Environment, an agent-based multimedia learning environment (see Figure 1). While interacting with the environment, participants were prompted to make metacognitive judgments (ease of learning judgments [EOLs], CEs, RCJs), and respond to multiple-choice questions, while learning about 9 different human body systems with multimedia materials. Participants interacted with the environment for 18 identical, self-paced trials.
Text and diagram CEs were coded based on their accuracy, while their responses to the multiple-choice questions were coded by correctness. Lastly, participants’ RCJs were collected on a scale from 50% to 100%.
Results from multi-level modeling (Raudenbush & Bryk, 2002) indicated that participants’ text CE accuracy (OR = 1.98, t = 3.09, p = .002) but not diagram CE accuracy (OR = 0.98, t = –0.10, p > .05) was associated with a 98% increased chance of responding to the multiple-choice question correctly. Additionally, results revealed that an increase in text CE accuracy (γ = 5.70, t = 3.95, p < .001) and diagram CE accuracy (γ20 = 6.01, t = 4.63, p < .001) significantly predicted an increase in participants’ RCJs, such that participants were more confident in their multiple-choice responses when they had more accurate diagram CEs. A significant interaction was found between participants’ diagram CE accuracy and multiple-choice responses (γ40 = –7.21, t = –2.75, p = .006), such that participants whose diagram CEs were accurate and who had more accurate multiple-choice responses also reported more confidence in their answers (see Figure 2).

Significance: Results from this study indicate accurate metacognitive judgments are required for successful multimedia learning. Traditionally, metacognitive judgments during multimedia learning have been found to be largely inaccurate. Compared to judgments traditionally examined during multimedia learning, our results indicate other metacognitive judgments may be more informative of successful learning with multimedia.

Authors

• Nicholas Vincent Mudrick, North Carolina State University

• Roger Azevedo, North Carolina State University

• Michelle Taub, North Carolina State University