Using real-time trace data to predict collaboration quality and creative fluency in design teams

Abstract

In this study, sixteen Engineering students were assigned to small groups (n=4) to work collaboratively on engineering design tasks. Using wearable sociometric devices, we collected real-time non-linguistic speech data on team interaction including turn-taking, successful interrupts and overlaps. Results from 2-stage regression models indicate that speech and conversational dynamics such as turn-taking and successful interrupts are significant in predicting the perceived collaboration quality and creative fluency of design teams.