Learning analytics with learning and analytics : advancing student agency analytics

Abstract

Pedagogically meaningful, research-based, and ethical learning analytics could foster the values and learning aims we want to advance in our society and educational system. However, it is essential to combine knowledge of the learning sciences and computational sciences when developing and applying learning analytics. This dissertation advances ananalytics approach called student agency analytics that utilizes learning analytics methods and computational psychometrics. Student agency is a vital characteristic of a learner, especially during times of uncertainty and change. Student agency has been raised to an important position in educational policymaking, and it has been identified as an essential aspect to consider when facilitating lifelong learning. The research advances the analysis process, examines the results from the student and teacher point of view, and provides novel insights into student agency. Specifically, the research addresses the issue of how to combine theoretical knowledge of learning and analytical methods as a comprehensive process in learning analytics while taking into account teachers’ perspectives, methodological issues, and some limitations in learning analytics. The results show that i) student agency can be characterized, and different profiles can be generated using robust clustering, ii) higher course satisfaction and performance is associated with higher student agency, iii) students reporting low agentic resources experience various restrictive aspects in learning, iv) explainable artificial intelligence techniques can provide additional insight about the intricacies of student agency, and v) teachers can utilize the analytics results in professional reflection and pedagogical decision-making.