Predicting Academic Performance: A Comprehensive Electrodermal Activity Study

Electrodermal Activity (EDA) provides insights into the Sympathetic Nervous System (SNS) activation, primarily responsible for “fight-or-flight” response, including stress. Stress, in turn, significantly affects academic performance. Despite EDA’s widespread application in various fields, its use in educational research remains underexplored. Existing studies on EDA’s relationship with academic outcomes often focus on isolated aspects of the signal, thus narrowing their scope and findings. This paper presents a comprehensive analysis of the entire EDA signal, employing it as input for regression models for predicting students’ grades. By augmenting the dataset to compensate for sample size limitations, we achieved low prediction errors (MSE of 0.002 and RMSE of 0.045) and high explanatory power (up to 80% of \(R^2\) ). Our findings indicate a preference for regression models capable of capturing non-linear relationships with minimal complexity. This research underscores the potential of using EDA as a tool for educators to identify students whose academic performance may be adversely affected by stress, paving the way for targeted interventions.