Student engagement and academic performance in pandemic-driven online teaching: An exploratory and machine learning approach

Abstract

Fostering student engagement to acquire knowledge and achieve academic performance requires understanding how students engage in learning and its influence on academic achievement. This provides valuable insights that help improve learning experiences and outcomes. The paper relies on a mixed methods approach by expanding the traditional dimensions of student engagement and by employing a machine learning framework to identify which specific dimension of student engagement exhibits the main impact on student academic achievement. A questionnaire-based survey is conducted for the period 2020-2021 among a cohort of Romanian students. The outcomes of this preliminary exploratory analysis are further embedded into a machine learning framework by performing a LASSO regression. The findings reveal that the most relevant dimensions of student engagement, during remote education, that contribute the most to outcomes were represented by the behavioural, social, cognitive, and emotional engagement dimensions. Furthermore, the switch to online education appeared to have inverted the positive relationship between social and cognitive engagement and academic achievement. Despite the inherent challenges, the student’s interest in class participation and homework completion was stimulated, and they managed to adapt without difficulty to study independently.

Keywords

• student engagement
• academic achievement
• remote education
• linear regression
• machine learning