Please use this identifier to cite or link to this item:
Title: Predicting Idea Co-Construction in Speech Data using Insights from Sociolinguistics
Authors: Gweon, Gahgene
Jain, Mahaveer
McDonough, John
Raj, Bhiksha
Rose, Carolyn
Issue Date: Jul-2012
Publisher: International Society of the Learning Sciences (ISLS)
Citation: Gweon, G., Jain, M., McDonough, J., Raj, B., & Rose, C. (2012). Predicting Idea Co-Construction in Speech Data using Insights from Sociolinguistics. In van Aalst, J., Thompson, K., Jacobson, M. J., & Reimann, P. (Eds.), The Future of Learning: Proceedings of the 10th International Conference of the Learning Sciences (ICLS 2012) – Volume 1, Full Papers (pp. 435-442). Sydney, NSW, AUSTRALIA: International Society of the Learning Sciences.
Abstract: Automatic assessment of group processes in collaborative groups is one of the holy grails of the computer supported collaborative learning community. In this paper, we present work towards detecting one type of group process which provides an important window into the inner workings of a group, namely "idea co-construction (ICC)". What is unique about our approach in relation to other educational data mining techniques is that we adopt insights from sociolinguistic theories by modeling stylistic convergence of speech. We present an unsupervised machine learning technique that is able to generate a predictor of the prevalence of ICC in face-to-face debates with an R2 value of .13, p < .05.
Appears in Collections:ICLS 2012

Files in This Item:
File SizeFormat 
435-442.pdf240.64 kBAdobe PDFView/Open

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.