Watch What You Say: Using Speech Processing to Promote Student Discussion in the Classroom

• In Event: Roundtable Session 55
  In Roundtable Session: Moving Away From Technology for Technology's Sake: Effective Practices for Student and Teacher Learning

Abstract

Classroom discourse is crucial for teaching and learning, but we lack practical ways to provide teachers with feedback on their classroom conversations. We describe an automated system to track the amount of discussion for elementary school mathematics lessons. The system uses the LENA device to collect speech data. Subsequent classification analysis identifies classroom discussion and gives the teacher daily feedback. With daily feedback and professional development, teachers increased discussion time by about one-third across 20 lessons. The amount of discussion was related to instructional strategies. Automated speech processing proves promising to study and improve classroom instruction.

Authors

• Zuowei Wang, Educational Testing Service

• Kevin F. Miller, University of Michigan - Ann Arbor

• Kai S. Cortina, University of Michigan