The Promises and Limitations of Computer-Based Analyses of Classroom Discourse: A Study of Literature Discussions

• In Event: 21.062 - MTCC Roundtable Session Six
  In Roundtable Session: 21.062-4 - New Methods: Discourse, Narrative, and the Politics of Evidence

Abstract

We present the results of a study that blended qualitative methods and computer science to study student talk in literature discussions in 19 high school English classrooms. Our aims were to (1) study the frequency of four features of student talk grounded in current research (argumentation, specificity, text/world focus, and collaboration), (2) determine relationships between these features and discussion quality, and (3) build computer-based programs to detect these features. We found that features of student talk that predicted discussion quality were high specificity, text-focus, and challenges to others’ ideas. Next, we created a computer system that reliably detected the specificity of students’ utterances. In conclusion, we discuss the promises, limitations, and current methods for computer-based analyses of classroom discourse.

Authors

• Amanda J. Godley, University of Pittsburgh

• Christopher Alan Olshefski, University of Pittsburgh