Please use this identifier to cite or link to this item: https://repository.isls.org//handle/1/145
Title: Local Versus Global Connection Making in Discourse
Authors: Collier, Wesley
Ruis, Andrew
Shaffer, David Williamson
Issue Date: Jul-2016
Publisher: Singapore: International Society of the Learning Sciences
Citation: Collier, W., Ruis, A., & Shaffer, D. W. (2016). Local Versus Global Connection Making in Discourse In Looi, C. K., Polman, J. L., Cress, U., and Reimann, P. (Eds.). Transforming Learning, Empowering Learners: The International Conference of the Learning Sciences (ICLS) 2016, Volume 1. Singapore: International Society of the Learning Sciences.
Abstract: This paper examines techniques for modeling relationships among domain concepts and practices in discourse to assess learning in a CSCL environment. We compare two approaches: a traditional psychometric approach, which models the global correlation structure of student discourse markers across the learning intervention, and a model that accounts for the local correlation structure of discourse markers within activities. We investigate whether: (a) analysis of local correlation structure can identify significant differences between novices and relative experts; (b) these differences reflect meaningful differences the discourse; and (c) analysis of global correlation structure can identify significant differences between novices and relative experts. We assess whether an approach that models local relationships among concepts in a domain provides useful information beyond what might be extracted from a more traditional modeling approach. Our results indicate that techniques that account for local correlation structure can identify patterns in discourse not reflected in global correlation structure.
URI: https://repository.isls.org/handle/1/145
https://dx.doi.org/10.22318/icls2016.56
Appears in Collections:ICSL 2016

Files in This Item:
File SizeFormat 
56.pdf401 kBAdobe PDFView/Open


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