Using Feature Engineering from Online Learning Environments to Observe Social and Emotional Skills and Academic Performance

A focus of learning analytics research has been the collection, analysis, and interpretation of data from online academic technologies, particularly in higher education where these technologies are incorporated extensively into the student learning experience. In particular, the learning management system (LMS) has emerged as a technology frequently integrated into all courses, whether online, in person, or hybrid, with three-quarters of students using the LMS in all or nearly all of their courses. Despite the ultimate goal of improving student learning (vs. simple prediction), there are few empirical studies of intervention efforts. Student interactions with learning management systems (LMS) and other educational technologies provide detailed information about how students interact with learning resources and activities. In this study, the authors coupled social and emotional learning and learning analytics in four undergraduate courses making extensive use of the LMS and examined relationships among social and emotional (SE) skills, behaviors recorded in the LMS, and course grades. The findings showed that SE skills, LMS data, and grades are largely associated with one another in expected ways, with course grades being robustly correlated with LMS behaviors across all courses, while different sets of LMS behaviors correlated significantly with different SE skills. Implications of results, limitations, and future work are discussed. [This project was facilitated through a partnership with the University of Maryland, Baltimore County.]