학습분석학 기반 대학생의 학업저성취 예측 요인 탐색

Abstract

Objectives The purpose of this study is to explore the possibility of preventing academic underachievers by utilizing the data of enrolled students accumulated in the school academic system from the perspective of learning analytics. In particular, by exploring predictive factors for academic underachievement based on academic data before the start of the semester, we intend to predict and select risk groups for academic underachievement early before the start of the semester. The results of early prediction and screening are intended to be used to prepare a preemptive support plan for teaching and learning for those at risk of academic underachievement.
 Methods 670 students of 4-year A university located in the metropolitan area were selected as the research subjects and the academic data accumulated in the school academic system was analyzed. The independent variables are the previous semester's GPA, total credits completed, total number of leave of absence, total number of academic warnings, total number of F credit courses, and major course registration credits and liberal arts course registration credits for which course registration has been confirmed for the current semester. Dependent variables are GPA and low academic achievement in the current semester. For data analysis, SPSS 28.0 statistical program was used, linear regression analysis was performed to confirm the explanatory power of independent variables for grades, and logistic regression analysis was performed to derive predictive factors for low academic achievement.
 Results As variables predicting grades according to Research Question 1, the GPA of the previous semester, total credits completed, credits for major course registration in the current semester, and credits for general course registration in the current semester had a positive effect on grades. And the total number of F-credit courses and the total number of leave of absence were confirmed as predictors that negatively affected grades. According to Research Question 2, the variables predicting low academic achievement are GPA of the previous semester, total credits completed, credits for major course registration in the current semester, and credits for general course registration in the current semester have a negative effect on low academic achievement. And the total number of F grade courses, the total number of leave of absence, and the total number of academic warnings were confirmed as predictive variables that had a positive effect on academic underachievement.
 Conclusions This study has great educational significance in that it derives predictive factors for grades and academic underachievement based on the academic data of students enrolled in the school and prepares a method for teaching-learning support. The results of this study will be available as practical examples and reference materials for early prediction and support of grade and academic underachievers.