Causal Inference of Factors Affecting Students’ Academic Performance

Abstract

In this study, we apply causal inference methods to analyze the impact of various factors on the academic performance of university students. Using a survey-based dataset, we treat the students’ GPA as the outcome variable and explore how behaviors such as the use of ChatGPT, participation in extra learning, and usage of external learning management systems (LMS) influence academic results. We employ Propensity Score Matching (PSM) and regression techniques to estimate the Average Treatment Effect (ATE) of selected variables. Additionally, we estimate the Average Treatment Effect (ATE) of selected variables using propensity score matching and validate the robustness of our findings with refutation tests. The results highlight that certain technology-related behaviors, particularly external LMS usage and discipline-related study efforts, show significant causal relationships with academic outcomes. We also perform refutation tests to validate the robustness of our findings. This research contributes empirical evidence for educational technology strategies and offers practical insights for improving student learning outcomes in higher education.