Learning analytics and the Universal Design for Learning (UDL): A clustering approach

Abstract

In the context of inclusive education, Universal Design for Learning (UDL) is a framework used worldwide to create learning opportunities accessible to all learners. While much research focused on the design and students' perceptions of UDL-based learning settings, studies on students’ usage patterns in UDL-guided elements, particularly in digital environments, are still scarce. Therefore, we analyze and cluster the usage patterns of 9th and 10th graders in a web-based learning platform called I₃Learn.

The platform focuses on chemistry learning, and UDL principles guide its design. We collected the temporal usage patterns of UDL-guided elements of 384 learners in detailed log files. The collected data includes the time spent using video and/or text as a source of information, working on learning tasks with or without help and working on self-assessments. We used Exploratory Factor Analysis (EFA) to identify relevant factors in the observed usage behaviors. Based on the factor loadings, we extracted features for k-means clustering and named the resulting groups based on their usage patterns and learner characteristics. The EFA revealed four factors suggesting that learners remain consistent in selecting UDL-guided elements that require a decision (video or text, tasks with or without help). Based on these four factors, the cluster analysis identifies six different groups. We discuss these results as a starting point to provide individualized learning support through further artificial intelligence applications and inform educators about learner activity through a dashboard.