Using Sequential Pattern Mining to Uncover Scaffolding Patterns in Middle School Online Math Instruction

• In Event: AERA Roundtable Session Thursday 9:50 am Four Seasons Ballroom 2-3
  In Roundtable Session: Learning Analytics and Student Engagement (Table 1)

Abstract

This study examined scaffolding strategies used by teachers in an online math education platform for middle schoolers. We employed a novel methodology combining qualitative coding, machine learning, and sequential pattern mining to analyze large-scale discussion forum data. Findings showed teachers primarily use feeding back, instructing, and questioning as scaffolding strategies, adapting their approach based on student actions in asynchronous online environments. The research contributes to understanding effective online math instruction by revealing the contingent nature of scaffolding strategies. Our innovative methodology bridges large-scale data with nuanced understanding of math scaffolding dynamics. The findings have significant implications for teacher training, platform design, and development of AI-powered scaffolding agents in online mathematics education for middle school.

Authors

• Wenting Zou, Pennsylvania State University

• Feiwen Xiao, University of Pennsylvania

• Jiaju Lin, Pennsylvania State University

• Huanying Song, Pennsylvania State University

• Chenglu Li, University of Utah