The Obligation to Act: A Critical Historical Portrait of the Learning Analytics Landscape

This article provides a critical historical portrait of the learning analytics (LA) landscape. It defines LA as the collection, analysis, interpretation, and communication of data about learners to provide theoretically relevant and actionable insights. The review highlights successful applications of Early Warning Systems (EWS), including Purdue University's Course Signals and Wisconsin's Dropout Early Warning System (DEWS), which demonstrate LA's potential to improve student retention and graduation rates. However, the paper identifies significant evidence gaps, noting a lack of causal research proving that LA systems directly enhance student achievement. Technical challenges are discussed, specifically the loss of information when continuous risk variables are discretized into binary categories. Ethical concerns involve algorithmic bias, the potential for self-fulfilling prophecies, and the need for informed consent. The synthesis proposes moving beyond prediction by integrating LA with Multi-Tiered Systems of Support (MTSS). Mathematics education is identified as a prime area for LA application due to the hierarchical nature of mathematical competencies. The article concludes that while technology facilitates data processing, teachers remain the vital agents of change in implementing responsive pedagogical interventions.