Toward Equitable Mathematics Education: Exploring AI-Human Collaboration in Belonging-Centered Instruction Analysis (Poster 11)

• In Event: AERA Poster Session Wednesday 10:50 a.m.
  In Poster Session: Motivation and Belonging in Mathematics Education

Abstract

Addressing the need for efficient, scalable methods in equity-focused educational research, this study examines the integration of generative AI within the Computer-Assisted Qualitative Data Analysis (CAQDAS) framework. We developed an AI coding assistant using Claude 3.5 Sonnet to enhance Belonging-Centered Instruction (BCI) analysis in mathematics classrooms. Comparing AI and human coding across two 8th-grade mathematics videos, results show AI's potential for rapid processing (83% faster than humans) but reveal challenges in nuanced interpretation. AI coding assistant precision varied (55-66%) across classroom contexts, while human coders maintained 90%+ precision. This research contributes to evolving AI-assisted qualitative analysis in education, offering insights into hybrid approaches that leverage both AI capabilities and human expertise to advance equity in mathematics education.

Authors

• Jingyun Wu, Indiana University

• Hao Ke, Tsinghua University

• Jamaal Matthews, University of Michigan