Exploring Approaches for Capturing the Emergence of Complex Self-regulated Learning Processes

• In Event: Applying Complexity Frameworks to Self-Regulated Learning Research: From Theoretical Models to Practical Tools

Abstract

Self-regulated learning (SRL) is essential to efficiently learn with advanced learning technologies where learners set goals, apply cognitive and metacognitive strategies (e.g., evaluating content, summarizing, monitoring progress towards goals), and then modify those strategies to achieve optimal learning outcomes[17]. Empirical SRL research tends to focus on how and when specific strategies are used rather than capturing and measuring the overall emergence of SRL as learners enact those strategies. This approach limits our understanding of how SRL can be applied across different contexts and domains. The current paper addresses this gap by incorporating Complex Systems Theory (CST)[18] as a lens through which learners’ SRL can be interpreted, offering analytical tools and approaches that can be used to quantify the emergence of SRL.
SRL is a complex system in which system behaviors are self-organized, interaction dominant, and emergent[5]. To study SRL as a complex system, researchers can leverage CST concepts, such as far-from equilibrium, and analytical approaches, such as auto-recurrence quantification analysis (aRQA), to fully understand how learners demonstrate functional SRL behavior[19-20]. Functional system behaviors are characterized by balanced oscillations between stable and chaotic states whereas dysfunctional systems remain in unbalanced states of either stability or chaos[20]. Analytical approaches, such as aRQA, quantify identify nonlinear shifts in behavior, quantify repetitive patterns, and visualize shifts in these behaviors using recurrence plots. However, issues with this approach arise as aRQA metrics are typically provided after the task has occurred and do not update in real-time. This presentation will address gaps in literature by describing and providing an empirical use case for two analytical approaches for capturing the emergence and fluctuation of learners’ (dys)functional SRL processes during the learning process.
The two analytical approaches discussed in this presentation are cumulative aRQA and binned aRQA. Cumulative aRQA constantly re-calculates learners’ aRQA metrics with each data added to a time series, resulting in cumulative metrics like students’ grade point average. With cumulative metrics, researchers can consider the full history of the learner when identifying (dys)functionality of their SRL processes where researchers can then identify the rapidity in which the system changes in terms of (dys)functionality. Binned aRQA calculates aRQA metrics over time according to events or phases, segmenting learners’ time series. For each bin, new aRQA metrics can be calculated and compared, allowing researchers to understand how learners’ SRL functionality changes according to events that have occurred within the environment or as a result of the phase of SRL they enacted.
These analytical approaches open the door for researchers’ interpretation of what “effective” SRL behavior is, how it can be quantified, and the manner in which SRL functionality can be supported via scaffolds embedded in advanced learning technologies. This presentation will conclude with a discussion on the research questions that can be addressed, the advantages and disadvantages of these analytical approaches, and the future of educational research in applying these methods to not just SRL, but other psychological constructs.

Authors

• Daryn Dever, University of Florida

• Megan Wiedbusch, University of Central Florida

• Roger Azevedo, University of Central Florida