A Psychological Early-Warning Algorithm for Math Anxiety Risk Using Multimodal Data
Xudong Sha, Li ZhaoAbstract
Objective
This study aims to develop and validate a psychological early-warning algorithm for math anxiety risk using multimodal data. The research seeks to identify at-risk students through integrated analysis of physiological, behavioral, and performance indicators, enabling timely intervention to prevent the development of severe math anxiety.
Subjects and Methods
The study involved 300 middle school students who completed math anxiety assessments (MAS) and participated in computer-based mathematical tasks while multimodal data were collected. Physiological measures included heart rate variability and electrodermal activity; behavioral data comprised eye-tracking patterns and response times; performance metrics included accuracy and problem-solving strategies. Machine learning algorithms (including SVM and Random Forest) were trained to identify math anxiety risk based on the multimodal features.
Results
The multimodal algorithm achieved 87.3% accuracy in predicting math anxiety risk, significantly outperforming single-modality approaches (70.1-78.5%). Key predictive features included prolonged fixation duration on difficult problems (p < 0.01), increased skin conductance response (p < 0.05), and specific error patterns in problem-solving. The algorithm successfully identified at-risk students 4-6 weeks before traditional screening methods, with a sensitivity of 85.2% and specificity of 89.1%.
Conclusions
The multimodal early-warning algorithm provides an effective tool for proactive identification of math anxiety risk, enabling timely psychological support. This approach demonstrates the value of integrating multiple data sources for mental health risk assessment in educational settings and offers a promising direction for preventing the negative impact of math anxiety on academic achievement and psychological well-being.
Corresponding Author
Xudong Sha, Department of Mathematics and Physics, Zibo Normal College, Zibo, Shandong, China.