Impact of learning analytics-based personalized feedback: an experimental study on student behavior in a large introductory STEM course

Abstract This study examined the impact of learning analytics-based personalized feedback on student engagement behavior, learning outcomes, and perceptions of the feedback in a large introductory STEM course. Students were randomly assigned to either an experimental group, which received four personalized feedback messages between Weeks 6 and 14, or a control group, which received generic messages consistent with the instructor’s previous teaching practices. The feedback was personalized based on students’ engagement with homework and practice problems and was designed to promote specific behaviors in these two areas. Results showed that the experimental group exhibited greater engagement in the targeted behaviors compared to the control group. However, no significant differences were found between the groups in perceived helpfulness of the feedback or learning outcomes. This study addresses the limitations of the current implementation of personalized feedback and highlights the value of integrating behavioral data with learning outcomes and self-report responses to more accurately evaluate its impact.