Detecting Self-Regulated Learning, Metacognition, and Emotions With Game-Based Learning Environments

• In Event: Advances in Self-Regulated Learning Research Using Learning Technologies: A Synthesis of the Empirical Evidence

Abstract

Objectives and Theoretical Framework. Game-based learning environments (GBLEs) are an effective means for fostering student learning, as research shows they have led to increases in student motivation, engagement, emotions, and self-regulation (Clark et al., 2016; Plass et al., 2020). The information processing theory of self-regulated learning (Winne, 2018) posits there are cognitive internal factors (e.g., emotions) that can impact how a student engages in different phases of learning. The model of affective dynamics (D’Mello & Graesser, 2012) states that if a student is faced with an impasse during learning, this causes a state of confusion, which if not resolved, results in frustration.
Although these processes impact overall learning, less is known about the temporal dynamic between emotions and the use of self-regulatory processes during learning. Therefore, the goal of this study is to examine the relationship between learner-centered emotions (confusion and frustration) and in-game self-regulatory activities during learning with CRYSTAL ISLAND, a GBLE that foster the use of self-regulated learning and scientific reasoning during science learning (Rowe et al., 2011; Taub et al., 2020).
Methods, Data Sources, and Results. Participants were college students from a large university in the U.S. who played CRYSTAL ISLAND (see Figure 1). The goal of the game is to solve the scientific mystery of what illness is impacting island inhabitants. To solve this mystery, students are required to gather clues by navigating to different locations on the island, form hypotheses, and test those hypotheses, which can all be done by engaging in different gameplay activities (e.g., reading book content, talking to virtual characters, testing food items, updating a monitoring tool). For this study we categorized gameplay activities as: (1) knowledge acquisition, (2) help seeking, (3) scientific reasoning, and (4) metacognitive monitoring. To successfully solve the mystery, students were required to correctly identify the illness and transmission source, and suggest a correct treatment plan. During gameplay, we collected multimodal data, including students’ facial expressions and log-file data.
Results from mixed effects models revealed significant interactions during metacognitive monitoring (see Figure 2): students with high evidence scores of confusion at the onset of metacognitive monitoring and a short duration of confusion during that metacognitive monitoring instance had the lowest evidence scores of confusion at the end of that instance. For frustration, students with the lowest evidence scores at the end of the metacognitive monitoring instance had a high evidence score at the onset with a long duration.
Scholarly Significance. These results demonstrate the differential impacts of confusion and frustration when engaging in self-regulatory activities during game-based learning, demonstrating the complex relationship between emotional and metacognitive processes. Specifically, we found significant interactions between both confusion and frustration when using the metacognitive monitoring tool, and the lowest evidence of each emotion at the end of that instance was associated with a high level at the beginning, but with different durations. This demonstrates that theoretically, the temporality and duration of each emotion plays a role in a student’s self-regulated learning behaviors during game-based learning.

Author

• Michelle Taub, University of Central Florida