SRL
Components Predicting Math Results in Estonian Middle School: A Bayesian Analysis
Elina Malleus‐Kotšegarov, Eve Kikas, Kati Aus, Danial Hooshyar ABSTRACT
Self‐regulated learning (SRL) components, such as motivation, cognitive skills and learning strategies, play a critical role in mathematics (math) learning. Using Bayesian Networks (BN), this study examines which SRL components predict membership in high (75th percentile) and low‐performing (25th percentile) groups in factual/procedural and conceptual math among middle school students ( N = 658). The results revealed that cognitive skills—specifically identifying key information in texts, scientific thinking and attention—were direct predictors of factual/procedural and conceptual knowledge. Self‐efficacy also emerged as a consistent motivational predictor across all models. Quick and frequent help‐seeking during tasks was found to predict belonging to a group of low conceptual knowledge, while text comprehension and working memory were specifically related to factual/procedural skills. Indirect influences among SRL components were also identified. The study highlights how different SRL variables contribute to various dimensions of mathematical competence and contrasts the most influential predictors of belonging to either high‐ or low‐achieving student groups.