Cross-course generalizability of SRL-aligned predictive models using digital learning traces

2026/04/30 by Jakob Schwerter, Loreen Sabel, Judith Bose +5 · 1 voice
Computer Science · Physics and Astronomy · Social Sciences · #Model Reduction and Neural Networks #Model-Driven Software Engineering Techniques #Student Assessment and Feedback

paper · doi:10.1016/j.compedu.2026.105670

openalex publication_date 2026/05/26 · openalex created_date 2026/05/27 · openalex updated_date 2026/06/12

Abstract

• SRL-aligned digital trace indicators enable early prediction of exam risk in theory-intensive computer science courses. • Predictive models generalize only partially across courses and institutions, with substantial performance loss under base-rate shifts. • Elastic Net models generalize more robustly than high-performing ensemble methods despite lower in-sample accuracy. • Probability calibration improves the interpretability of predicted risk but cannot fully compensate for cross-institution base-rate differences. • Stable SRL-related behavior patterns predict success even when their week-level manifestations vary.

Citations

• Learning Analytics
• Predicting student outcomes using digital logs of learning behaviors: Review, current standards, and suggestions for future work
• Revisiting the Metacognitive and Affective Model of Self-Regulated Learning: Origins, Development, and Future Directions
• Using multimodal learning analytics to validate digital traces of self-regulated learning in a laboratory study and predict performance in undergraduate courses.
• Is Big Data Better? LMS Data and Prediction Accuracy in Postsecondary Education
• Interpreting Predictive Learning Sequences in a College Math Course through a Self-Regulated Learning Framework
• Which Imputation Fits Which Feature Selection Method? A Survey-Based Simulation Study
• Variable selection with multiply-imputed datasets: choosing between stacked and grouped methods
• Variable Selection with Multiply-Imputed Datasets: Choosing Between Stacked and Grouped Methods
• State-of-the-art in selection of variables and functional forms in\n multivariable analysis -- outstanding issues
• Attaining Self-Regulation

Discussions

• Important work on whether predictive models can be ported across courses and institutions. The greater the differences in course design and/or base-rate of student difficulty, the less the models port [bsky, 1 points, 0 comments]

Related

• Interpreting Predictive Learning Sequences in a College Math Course through a Self-Regulated Learning Framework
• Toward Cyclic A.I. Modelling of Self-Regulated Learning: A Case Study with E-Learning Trace Data
• Focus on Likely Classes for Test-Time Prediction
• Prediction of Success or Failure for Final Examination using Nearest Neighbor Method to the Trend of Weekly Online Testing
• Feature Engineering on LMS Data to Optimize Student Performance Prediction