Rhetorical structure analysis for assessing collaborative processes in CSCL

The purpose of this paper is to present a research on using the rhetorical structure analysis for assessing collaborative processes in Computer-Supported Collaborative Learning (CSCL) chats. The collaborative process is of great importance in CSCL research, especially in the case of using chat which is one of the most important tools in CSCL. Hence, it is important to assess collaboration in chats, which presumes identifying the intertwining of ideas or points of view among participants. For this purpose, a specific model is assumed to analyze and assess the collaboration in chat. This model depends primarily on the ideas of Mikhail Bakhtin's dialogism theory and Stefan Trausan-Matu's polyphonic model in order to extract the threads (repeated words) discussed in the chat. Moreover, cue phrases and their usage in linking threads are also identified in order to detect the collaboration between participants. For this purpose, integrated tools were used with comprehensive functionality (Natural Language Processing techniques, linkage process and chat analysis) in order to analyze the content of the chat. The results reveal that the toolkit we developed based on the assumed model is quite effective in helping teachers to analyze and assess the students' collaborative chats. The results are presented concisely in statistical tables and graphics to make our understanding of the collaborative process easier. It also allows students to know and understand the interactions and how it contributes to the conversation.