Towards Triggering Adaptive Collaborative Learning Support Using Automated Conversational Analysis
Carolyn R. Rosé
Language Technologies Institute and the Human-Computer Interaction Institute, Carnegie Mellon University
Newell-Simon Hall 1305 (Michael Mauldin Auditorium)
Our recent text classification work (Donmez et al., 2005; Wang et al., submitted) has produced technology capable of real-time analysis of collaborative learning discussions, which opens up the new possibility for adaptive collaborative learning support in the midst of free form on-line communication. Results from wizard-of-oz investigations of support in this form demonstrate promise that this technology can yield a positive learning benefit for collaborative learners (Gweon et al., 2006). Two studies in the past year evaluating a fully-automatic form of this collaborative learning support both demonstrate significant learning benefits for students in comparison with a no support control condition (Wang et al., submitted; Kumar et al., submitted).
In this talk, I will describe the motivation and long term goals of this research agenda. I will then describe an architecture for triggering fully automatic adaptive collaboration support based on an analysis of student conversational behavior, and how it can be adapted for different conversational analysis frameworks and alternative designs for collaboration support. I will briefly outline the series of results from the work to date on this topic. Finally, I will describe in detail one study we have conducted with this approach in a science inquiry learning task. In this study, we explore the effect of adaptive support on idea production and learning with a 2×2 factorial design, where we manipulate the presence or absence of adaptive prompts and whether students worked individually versus working in pairs as two between subjects factors. This experimental design allows us to evaluate both the direct effect of adaptive prompts on learning as well as the indirect effect they might have on learning through improving the quality of the collaboration process itself. In addition to the positive benefit of the support on learning, a detailed process analysis also demonstrates an impact of the adaptive support on idea generation productivity in pairs (Wang et al., submitted).
Carolyn Penstein Rosé joined the faculty at the Language Technologies Institute and the Human-Computer Interaction Institute in Fall of 2003. A particular focus of her research is in exploring the role of explanation and language communication in learning and in supporting productive learning interactions with language technologies.