What Makes Tasks Discussion-Worthy? A Computational Discourse Analysis

• In Event: Mathematics Tasks: Four Different Perspectives

Abstract

Mathematics task design does not determine how teachers and students ultimately take up activities. However, task features may encourage particular language practices like explaining why or engaging with others’ ideas. After analyzing 55 tasks’ features, I qualitatively coded transcripts from lessons where students did those tasks. In order to look for associations between task features and student discourse, I employed a computational language model to extrapolate codes to all 446 observations of student discussions. Findings suggest, when accounting for the influence of teachers, certain task design features are associated with differences in student discourse. I hope this project contributes to ongoing research on task design, language-responsive pedagogy, and methodologies that can expand what is possible with discourse analysis.

Author

• Jim Malamut, Stanford University