How does student behavior affect performance in a project-based course?

Zhang, H., An, M., Kandimalla, S., et al.

In this work, we focus on the relationship between student behavior, consistency in behavior, and performance in a project-based online course. To capture behavior, we collected log data from two semesters from the project portal, the auto-grading system, and a cloud service provider. Using the logs, we analyzed both direct and synthesized metrics associated with the learning process. We measure the consistency of student behavior by the variance of their behavior percentile across all projects. Correlation is generally not found between behavior and performance in the spring semester but found in the fall. Further, a statistically significant correlation between behavior and performance is found for the spring data, but only partially in the fall. By studying behavior and consistency in behavior across the fall and spring semesters, we can provide insights into how students develop work habits necessary in a demanding environment and potentially relates consistency to students’ academic maturity.

Hongyi Zhang, Language Technologies Institute
Marshall An, Computer Science Department
Siddharth Kandimalla, Computer Science Department
Seth Goldstein, Computer Science Department
Majd Sakr, Computer Science Department