The Mediating Role of Learning Analytics

Abstract

Researchers in learning analytics have created indicators with learners’ trace data as a proxy for studying learner behaviour in a college course. Student Approaches to Learning (SAL) is one of the theories used to explain these behaviours, distinguishing between deep, surface, and organized study. In Latin America, researchers have demonstrated that organized approaches to learning could be more effective in higher education, leading to better performance and course approval. However, further analysis of student behavioural data is needed to understand this relationship and inform interventions targeting study habits and academic performance. In this study, we analyzed the relationship between student approaches to learning and their final grade in six college courses, using behavioural trace data as a mediator variable. Specifically, we conducted a quantitative study in two Latin American institutions where data of different granularity was collected from their Learning Management Systems. We observed that most learning analytics indicators do not mediate the effect between approaches to learning and course performance. However, there was evidence for fine-grained indicators acting as total mediators. Implications are discussed at methodological and pedagogical levels, aiming to inform the advancement of learning analytics in the region and its use for supporting student learning.