Modeling Effective and Ineffective Knowledge Communication and Learning Discourses in CSCL with Hidden Markov Models

Research output: Contributions to collected editions/worksChapterpeer-review


The grounding theory according to Clark (1996; Clark and Brennan, 1991) is a prominent approach to describe the co-construction of knowledge in computersupported collaborative learning (CSCL) environments. Communication problems such as incoherence or inadequate coordination are common in simultaneous textbased chat tools resulting in impaired grounding processes, which again may affect learning outcomes. Thus collaboration scripts like learning protocols are implemented to increase the structure of chat discourses and to reduce communication problems in order to support grounding processes. In this study the impact of grounding itself on learning outcomes was examined by implementing a

confederate in each learning group, who showed either a constructive or a destructive grounding behavior. Probabilistic models were computed for both the discourses of the present study as well as effective and ineffective discourses of a previous study. The long-term objective of this research is an adaptive script automatically detecting discourse structures in order to support grounding and thus to facilitate knowledge communication and learning processes.
Original languageEnglish
Title of host publicationAdvances in Applied Digital Human Modeling
EditorsVincent G. Duffy
Number of pages10
Place of PublicationBoca Raton
PublisherCRC Press
Publication date01.01.2010
ISBN (Print)9781439835111
ISBN (Electronic)9780429151842
Publication statusPublished - 01.01.2010

    Research areas

  • Business psychology
  • Adaptive scripting, Chat, CSCL, Grounding, Hidden markov models