Understanding Math Literacy Gaps: A Cross-National Machine Learning Study of Motivational Predictors

• In Event: AERA Roundtable Session Sunday 7:45 am - Gold 3
  In Roundtable Session: International Perspectives on the Teaching and Learning of Mathematics

Abstract

Factors driving mathematical literacy across contrasting educational systems remain incompletely understood. Guided by the Self-System Model of Motivational Development (SSMMD) which posits that context, self-system processes, and action jointly shape outcomes, this study applied machine learning to PISA 2022 data from 20,045 15-year-old students (Singapore: n = 6,606; Indonesia: n = 13,439) to compare multi-level predictors of mathematical literacy across contrasting educational systems with the result shows Gradient Boosting Machine outperformed other models. Psychological factors (e.g., self-confidence) dominated the prediction of mathematical literacy in high-performing Singapore, whereas contextual factors (e.g., socioeconomic status) were more dominant in Indonesia. Cross-country differences in predictor strength suggest that the effectiveness of motivational constructs may depend on system-level coherence and performance.

Authors

• Yiran Lin, McGill University

• LUYAO XU, McGill University

• Nikki G. Lobczowski, McGill University

• Maria Cutumisu, McGill University