Temporal patterns and visualizations of peer talk: Toward understanding the process and performance of dialogic collaborative problem-solving

Abstract

Dialogic collaborative problem solving (CPS) describes how students collaboratively solve a problem mainly through talk. Existing studies intensively explored cumulative features of productive peer talk based on a coding-and-counting approach. Nevertheless, it has not been fully understood how utterances historically and dynamically unfold overtime and gradually shape the group solution quality. This dissertation aims to identify temporal patterns of peer talk that can distinguish high-performing groups from low-performing groups in the dialogic CPS and examine the degree to which these temporal patterns help better predict group solution quality.