What is it about?

Reflective activities in undergraduate classrooms was found to be effective in improving students' learning and motivation. In this paper, we leveraged LLMs to dig deeper into students' learning reflections, in order to better understand whether sentiments on students' learning reflections are associated with academic performance.

Featured Image

Why is it important?

Our findings show that the percentage of positive sentiments in reflections on what students found confusing in their learning is negatively correlated with their academic performance. This result aligns with prior research that found sentiments to be negatively correlated with some measures of academic performance. Future research can explore further on the possible implications of this negative correlation between sentiment and academic performance and develop strategies to effectively intervene and address these sentiments.

Perspectives

Conducting this research was memorable for me because it happened at a particularly difficult time of my life when I was struggling with significant health issue. This research started as a course project and I developed it into a research publication with the help of my professors and research group, and I am grateful for the journey as it allowed me to grow further as a researcher and member of the community. This paper is also a part of a larger exploration that I've done in my PhD journey in digging deeper into reflective activities in undergraduate classrooms. I am glad to contribute in improving our understanding on what we can identify from students' learning reflections, which is crucial to address their learning challenges and improve their learning achievements.

Alfa Satya Putra
Purdue University

Read the Original

This page is a summary of: WIP: Leveraging LLM for Sentiment Analysis of Student Reflection Texts from a Large Undergraduate Course, July 2025, ACM (Association for Computing Machinery),
DOI: 10.1145/3698205.3733937.
You can read the full text:

Read

Resources

Contributors

The following have contributed to this page