Analyzing Communal Conversation
College: Graduate School of Library and Information Sciences
Award year: 2007-2008
Interaction in email, listservs, bulletin boards, chat, and blogs is creating a growing legacy of texts describing the social activity of groups and communities, many of which support critical processes of societal and commercial impact. Yet there is little that helps us examine these transcripts in a standard, scientific way. At this time, neither the technical infrastructure of automated methods and tools, nor the social science that support their use and interpretation exist to help those creating, maintaining and/or managing online community interaction. This project brings together the existing expertise and technologies of NCSA -- notably the work of Michael Welge and Xavier Llorà on the DISCUS project (Distributed Innovation and Scallable Collaboration in Uncertain Settings project) -- with work done by Caroline Haythornthwaite and doctoral student Anatoliy Gruzd on examining online transcripts using natural language processing techniques. Our focus is on building a viable platform for social science analysis of online transcripts. The goal of this project is to build an analysis environment suitable for researchers and for key participants in the context that will be used as a test case for this work, that of discussion forums from a series of online classes. The focus is on methodological, theoretical and practical outcomes, including identifying and addressing technical issues in such analyses, developing and applying theoretical approaches for analysis of these data, and providing a practical system of interrogation of these transcripts that can be of use to instructors and students in such classes. Future plans include extending this platform to analysis of other online forums, and to comparative analyses of the character of such forums.