Analytical Approaches for Examining Learners’ Emerging Self-regulated Learning Complex Behaviors with an Intelligent Tutoring System

Abstract

Self-regulated learning (SRL), or the ability for a learner to monitor and change their cognitive, affective, metacognitive, and motivational processes, is a critical skill to enact, especially while learning about difficult topics within an intelligent tutoring system (ITS). Learners’ enactment of SRL behaviors during learning with ITSs has been extensively studied within the human-computer interaction field but few studies have examined the extent to which learners’ SRL behaviors quantitatively demonstrate a functional system (i.e., equilibrium of repetitive and novel behaviors). However, current analytical approaches do not evaluate how the functionality of learners’ SRL behaviors unfolds as time on task progresses. This paper reviews two analytical approaches, both based within categorical auto-recurrence quantification analysis (aRQA), for examining how learners’ SRL complex behaviors emerge during learning with an ITS. The first approach, binned categorical aRQA, segments learners’ SRL behaviors into bins and performs categorical aRQA on the SRL behaviors enacted within those bins to produce metrics of complexity that demonstrate how learners’ functionality of their SRL systems change over time. The second approach, cumulative categorical aRQA, continuously calculates complexity metrics as learners enact SRL behaviors to identify the evolution of learners’ functional SRL. These two approaches allow researchers to identify how the functionality of SRL behaviors change over time in relationship to the occurrences within the ITS environment. From this discussion, we provide actionable implications for contributing to how learners’ SRL functionality can be visualized and scaffolded during learning with an ITS.