Scholarlink

<p>This study analyzed EdNet and ASSISTments datasets (131,441 middle school students) to develop a behaviorally grounded psychological model of emotion-cognition interaction in mathematics learning. Using ensemble learning for the behavioral inference of emotional states (87.3% classification accuracy), Bayesian Knowledge Tracing, and Hidden Markov Models, four interaction patterns were identified: synergistic facilitation (35%), competitive inhibition (28%), emotion-dominant (20%), and cognition-dominant (17%). Emotion-cognition interaction explained 23.6% of learning performance variance, surpassing emotional (12.4%) or cognitive factors alone (15.8%). Temporal analysis revealed a U-shaped trajectory: interaction intensity peaked initially (β = 0.51), reached minimum mid-stage (β = 0.29), and rebounded during challenging tasks (β = 0.46). Significant individual differences emerged between high (coefficient = 0.42) and low emotional regulation groups (0.18), emphasizing metacognitive abilities’ role in modulating interactions.The model demonstrated robust cross-dataset validation (correlation = 0.89). These findings provide a theoretical framework for understanding mathematics learning psychological mechanisms and inform personalized educational interventions and intelligent tutoring system design. The research highlights that emotion-cognition interaction, rather than isolated factors, critically determines learning outcomes, with temporal dynamics and individual regulation abilities as key moderators.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Psychological Mechanisms of Emotion-Cognition Interactions in Secondary Mathematics Learning: An Empirical Study based on Large-Scale Educational Behavioral Data

  • 期刊论文

  • 发表日期：

    2026年7月28日

查看全文
  • Yicheng Wang

摘要

This study analyzed EdNet and ASSISTments datasets (131,441 middle school students) to develop a behaviorally grounded psychological model of emotion-cognition interaction in mathematics learning. Using ensemble learning for the behavioral inference of emotional states (87.3% classification accuracy), Bayesian Knowledge Tracing, and Hidden Markov Models, four interaction patterns were identified: synergistic facilitation (35%), competitive inhibition (28%), emotion-dominant (20%), and cognition-dominant (17%). Emotion-cognition interaction explained 23.6% of learning performance variance, surpassing emotional (12.4%) or cognitive factors alone (15.8%). Temporal analysis revealed a U-shaped trajectory: interaction intensity peaked initially (β = 0.51), reached minimum mid-stage (β = 0.29), and rebounded during challenging tasks (β = 0.46). Significant individual differences emerged between high (coefficient = 0.42) and low emotional regulation groups (0.18), emphasizing metacognitive abilities’ role in modulating interactions.The model demonstrated robust cross-dataset validation (correlation = 0.89). These findings provide a theoretical framework for understanding mathematics learning psychological mechanisms and inform personalized educational interventions and intelligent tutoring system design. The research highlights that emotion-cognition interaction, rather than isolated factors, critically determines learning outcomes, with temporal dynamics and individual regulation abilities as key moderators.