2025 ASEE Annual Conference & Exposition

This study explores the relationship between student sentiment on post-assessment reflections and academic performance in a flipped, mastery-based Statics course. Utilizing Natural Language Processing (NLP) techniques, we analyzed over 28,000 student comments collected from four semesters with each student contributing an average of 54 comments per semester. This data provides a rich and comprehensive source of qualitative data for each student. Sentiment analysis was applied to each written comment to categorize the tone as either positive, negative, or neutral. Following the assessments conducted biweekly, students submitted a reflection as part of the learning process, commenting on each part of their solution. These comments provide insights into their cognitive and emotional engagement with the course material. To better understand each student’s sentiment, we developed an approach to quantify their overall sentiment score for each assessment. This method enables us to track the evolution of each student’s tone throughout the semester.

Our study will determine how this evolution in student sentiment correlates with their final grade in the course, identifying whether emotional tone in reflections is linked to academic performance. These trends could provide insights into how students perceive and engage with their work and how it aligns with the course metrics. This study also highlights new opportunities for targeted interventions in the course. Through leveraging NLP and reflective exercises, instructors gain access to more detailed and individualized insights into class progress. This can foster a better understanding of the connection between student attitudes and performance, enabling more personalized feedback and tailored interventions that can improve learning outcomes.