Statistical Analysis and Regression to Predict Students' Academic Performance in Online Learning

Abstract

One significant challenge in online learning is that teachers face difficulties monitoring students' academic performance. If unresolved, this issue may lead students to fail to achieve optimal academic outcomes. Understanding student behavior that influences academic performance is essential in online learning settings. This study is based on a dataset limited to a single course, Linear Algebra, involving 147 students. While this provides focused insight, it also restricts the generalizability of the findings to other subjects or academic contexts. This research aims to investigate the impact of student behavior on academic performance during online learning, focusing on the duration of participation in lectures and the number of student interactions. A quantitative regression approach was utilized to analyze the relationship between student behaviors and academic performance. The independent variables were formulated as DurationAve, ParticipationAve, DurationTotal, and ParticipationTotal. Four regression models were developed and evaluated based on statistical metrics, including R-squared, adjusted R-squared, standard estimation error, p-value, MSE, RMSE, and MAE, to identify the optimal model. The analysis revealed that the optimal regression model is: FinalScore = 69.7877 + 0.1552 DurationAve + 1.6060 ParticipationAve. The model demonstrated superior performance across all statistical measures. Furthermore, the findings indicated that ParticipationAve (average participation) had a greater impact on students' academic performance than DurationAve (average duration), as evidenced by the higher coefficient for ParticipationAve. The results confirm that the average number of student participations significantly influences academic performance in online learning environments. These findings support the need to foster interactive engagement during online classes. Future studies should explore additional factors that influence academic outcomes, such as demographic information, prior academic performance, and external influences, to further enhance understanding of predictors of academic success in online learning contexts.