Examining Task Context in SRL using Distance Measures and Sequential Adherence

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

Abstract

Objectives & Theoretical Framework
High-structure active learning courses are designed to focus students’ effort and distribute learning activities around a lecture session[21-23]. This affords learners more opportunities to actively engage prior to, during, and after class time, as instructors integrate multiple activities and learners choose to adhere to this co-regulated engagement in a lesson cycle[24] or engage in self-regulate learning (SRL) by taking their own approach to learning. SRL is often described as a loosely sequenced cycle during which learners engage in unconscious, cognitive, metacognitive, and affective processes that shape learning choices and behaviors[25]. Each of these processes is further situated in a context defined by features of both the task and individual[26]. We observed the activities of learners in an introductory gateway STEM course at a research university with a large enrollment of first-semester and first-generation learners. Learning analytics offers the opportunity to leverage digital traces to examine the fine-grain detail of students’ behavioral choices that are theorized to represent motivation, metacognition, or SRL[27]. By leveraging the emerging potential of learning analytics, we were able to assess the completion, sequence, and temporality of learners’ engagement with pre-, in-, and post-class activities through the first semester exam.
Methods & Data
To model complexity across learners’ (N=518) adherence and SRL in lesson cycles, we used learning analytics to model the ways learners’ contextualized action sequences departed from “adherent active learning” to self-regulated learning along a set of distance metrics we engineered and applied to eight weekly lesson cycles. Our initial models have focused on learners’ sequential adherence. By adapting a d2 distance measure[28], we were able to calculate the distance of learners’ engagement in the tasks of a weekly lesson cycle from the order recommended by the course instructors. This distance measure was calculated for the eight lessons prior to Exam 1 and regressed on learners’ exam performance.
Results
We conducted a set of linear regression models to assess whether learners’ sequential adherence to the ordered lesson activity sequence was a significant predictor of performance. Initial findings indicate that sequential learning matters. Controlling for prior content knowledge, learners’ sequential adherence across lessons explained a significant amount of variance on Exam scores (∆R 2 = .035, F= 3.02, p < .004). Adherence in the eighth lesson, immediately before the exam, was a statistically detectable predictor of performance (Table 1).
Scholarly Significance
These initial results provide an argument for the importance of parsing the complexity of learning behaviors in a way that aligns to SRL theory[26], and evidence that engaging sequentially in the context of active learning lesson designs can impact learning. Temporality of action sequences may further influence their import, where sequential actions become more predictive in the wake of an impending exam. Additional investigations into temporality including timeliness, responsiveness, and procrastination are next to be explored discussed in the session.

Authors

• Michael Abdul Ghani Berro, University of North Carolina - Chapel Hill

• Robert D Plumley, University of North Carolina - Chapel Hill

• Shelbi Laura Kuhlmann, University of Memphis

• Laura Ott, University of North Carolina - Chapel Hill

• Kelly Hogan, Duke University

• Matthew L. Bernacki, University of North Carolina - Chapel Hill

• Jeff A. Greene, University of North Carolina - Chapel Hill