Publications

In this study, we applied a set of learning analytics methods (k-means clustering, data visualization) to clickstream data from an interactive online algebra game to unpack how middle-school students’ (N = 227) behavioral patterns (i.e., the number of problems completed, resetting problems, reattempting problems, pause time before first actions) correlated with their understanding of mathematical equivalence. 

The k-means cluster analysis identified four groups of students based on their behavioral patterns in the game: fast progressors, intermediate progressors, slow progressors, and slow-steady progressors. The results indicated that students in these clusters, with the exception of slow progressors, showed significant increases in their understanding of mathematical equivalence. In particular, slow-steady progressors, who reattempted the same problem more often than other students, showed the largest absolute learning gains, suggesting that behavioral engagement played a significant role in learning. With data visualizations, we presented evidence of variability in students’ approaches to problem solving in the game, providing future directions for investigating how differences in student behaviors impact learning. 

This study examines how student and instructor participation in online discussions impacts students’ course performance. The context for the study is university introductory online mathematics/statistics courses, which typically have much higher failure rates than their face-to-face counterparts. 

Using text-mining techniques, we analyze online discussion data automatically collected by a Learning Management System across five years from 2869 students in 72 online courses, who collectively contributed 20,884 posts. These semi-automated techniques enable a broader and more scalable view of participation behaviors by investigating: (1) student posting and non-posting behaviors (called online speaking and listening, respectively), (2) the textual content of posts, and (3) instructors’ strategies for structuring discussions.

 Multilevel modeling results show that online listening behaviors significantly predict students’ course performance. Further, students’ posts that built on other contributions or applied new knowledge have the highest predictive value in terms of course performance. Finally, the instructors’ use of open-ended prompts is the only variable positively and significantly links to students’ course performance. Links to theory, instructional practice, and educational data mining are discussed.