Analysis of Learning Behaviors in SPOC-Based Blended Teaching of Linear Algebra

Abstract

SPOC (Small Private Online Course), as a teaching model that deeply integrates online learning with classroom instruction, has been widely applied in the field of higher education in recent years. This study takes the Linear Algebra course at a certain university as an example and systematically analyzes the characteristics of students' learning behaviors based on log data from the SPOC blended teaching platform. The findings reveal: First, students' video-watching behavior exhibits a distinct "beginning-and-end concentration" pattern, with a negative correlation between video length and completion rate. Second, online test performance shows a significant positive correlation with final exam scores, but the submission times of chapter tests are considerably dispersed. Third, forum interaction displays a phenomenon of "marginal participation," with active posters accounting for less than 15% of students. Fourth, cluster analysis of learning behaviors categorizes students into three types: "balanced," "cramming," and "passive," with significant differences in academic performance among the different types. Fifth, overall satisfaction with blended teaching is high, but individual differences in student learning engagement are substantial. Based on these findings, this paper proposes improvement suggestions such as optimizing video resource design, strengthening process-oriented evaluation incentives, promoting deep forum interaction, and implementing stratified teaching interventions. The research conclusions provide data support and practical references for the effective implementation of SPOC blended teaching in mathematics courses.