RUME 27
2025
Alexandria, Virginia
Exploring Undergraduate Mathematics Students’ Math Identities Using Sentiment Analysis: Affordances and Precautions
Page: 527
Students’ identities, orientations, and experiences in mathematics courses shape their participation and longevity in the discipline. Yet, undergraduate mathematics instructors do not often integrate pedagogy or curricula that help them build knowledge of their students’ math identities. We believe that in many math courses this is a problem of time and scale. We propose that language processing methods can help alleviate this tension by giving instructors a baseline understanding of their students’ math identities. In this study, we applied sentiment analysis to 28 mathematics “origin stories” written by undergraduate pre-calculus students. We compared the predictions from four language models with qualitative codes applied by two mathematics education researchers. We conclude this paper by discussing the strengths and limitations of sentiment analysis in our data, and discuss implications for undergraduate mathematics researchers and practitioners who may wish to apply these methods to their work.