Implementing enriched context descriptions for efficient scaling of learning analytics

Abstract

Learning analytics (LA) collects data about learners and their contexts to enhance learning. Integrating data from multiple data sources may provide a more holistic picture of learners and enable more useful analysis than is possible from isolated data sources. Learning activity data standards, including xAPI, can support data integration and thus help scale up LA. Research, however, has shown such standards are not widely used for these purposes, possibly related to limits in their expressibility. In this study, we provide implementation of recommendations for enhancing xAPI context descriptions and expressibility, as part of a technical solution that could be used by users preparing activity data for LA. The technical solution, realized through the creation of an xAPI profile and following affordances and constraints of xAPI, is adapted to the K-12 school adaptivity case. The technical solution is evaluated through user testing with technical experts having real-world experience with xAPI data description, allowing participants to explore the solution at a general level and consider different aspects related to its implementation. The evaluation results indicate that the technical solution meets criteria for usefulness and effectiveness. Furthermore, the results indicate user satisfaction with the solution at a general level, related to the recommendations, while also pointing out room for improvement in terms of the implemented solution. By enhancing the context descriptions of xAPI, which enables Learning Analytics (LA) to analyze data from multiple sources as evidence and supports adaptation, this research has the potential to contribute to both adaptive learning and evidence-based practice.