Learning analytics and mathematics performance: the mediating role of student engagement and self-regulated learning

Abstract

Learning analytics is widely regarded as a tool for enhancing academic performance, yet empirical evidence on this relationship remains inconsistent, leaving it unclear whether learning analytics improves performance directly or through underlying psychopedagogical processes such as student engagement and self-regulated learning. This gap is especially evident in mathematics education, a cognitively demanding field in which the mediating roles of engagement and self-regulated learning have rarely been examined. This study therefore investigated how learning analytics affects mathematics performance, with student engagement and self-regulated learning as mediating variables. Data were collected from 126 undergraduate students enrolled in Learning Management System (LMS)-supported mathematics courses at Universitas PGRI Mpu Sindok, Indonesia, using a quantitative method with an explanatory survey design. We used questionnaires and essay-based math tests to collect data on higher-order thinking skills, then analysed the data using Partial Least Squares–Structural Equation Modelling. The results showed that learning analysis had a positive and significant effect on student engagement, with a path coefficient of 0.939. Student engagement and self-regulated learning significantly affected math performance, with coefficients of 0.494 and 0.359, respectively. However, learning analysis did not show significant direct effects on self-regulated learning or mathematical performance. The indirect influence of learning analysis on mathematics performance through student involvement was found to be significant, with a coefficient of 0.464. These findings highlight the significant role of student engagement in leveraging learning analytics to improve math performance.

The authors' abstract, as published at the source. Cogent Education, 2026 · DOI ↗