Online Interventions: Summarization and Conversational Agent Technology Triggered through Automatic Discourse Analysis

The major thrust of my post-tenure research has been focused on supporting collaborative and discussion based learning in MOOCs. With the recent press given to online education and increasing enrollment in massively open online courses, the need for scaling up quality computer-mediated educational experiences has never been so urgent. Current offerings provide excellent materials including video lectures, exercises, and some forms of discussion opportunities. The biggest limitations are related to the human side of effective educational experiences. This includes personal contact with instructors and the cohort experience. In the past decade, special concern has centered on students’ inability to communicate effectively, negotiate ideas, or engage in other aspects of collaborative problem-solving activity. These concerns prompt my research on interventions that improve the instructional value of online collaborative learning experiences.

At least a decade of research, including my own, shows that students can benefit from their interactions in learning groups when automated support is provided, especially interactive and context sensitive support. Until recently, the state-of-the-art in computer-supported collaborative learning has consisted of static forms of support, such as structured interfaces, prompts, and assignment of students to scripted roles. Now technology for dynamic support of collaborative learning is publically available. My group is widely recognized as playing a major role in enabling this paradigm shift, especially as a result of demonstrations that dynamic script based support for collaborative learning leads to improvements in learning over otherwise equivalent static forms of support, and leads to 1.24 standard deviations more learning than learning with the same materials alone and without the support of a conversational agent (Kumar et al., 2007). Its value has been recognized in award and award nominations for my group’s work at conferences such as ACM SIGCHI, AI in Education, the International Conference of the Learning Sciences, and Computer-Supported Collaborative Learning. In our work since tenure, we partnered with the Institute for Learning in launching a MOOC on Accountable Talk, with 60,000 enrolled students. Since then, we have deployed interventions to support collaborative learning in five different Massive Open Online Courses. In three completed deployments we have measured substantial reductions in attrition associated with participation in collaborative chat activities we have hosted in these contexts. In an ongoing MOOC deployment in collaboration with the Smithsonian Institute, we have deployed support to team project-based learning in the course.

At the time of my tenure review, the Bazaar architecture for support of collaborative learning activities had already been made publically available. Now we have added to that large scale infrastructure for data storage of discourse data and a customizable discussion forum plugin for the edX platform with easy integration of social recommendation interventions. We are in active partnership with edX, engaging in regular collaborative discussions about extensions of the platform for better support of collaborative and discussion based learning in MOOCs at a broad scale, with a plan for one of my PhD student to work side-by-side with edX engineers either over the Summer of 2016 or in Fall of 2016 on platform extensions. We are part of two Gates funded networks, including the Digital Learning Research Network, in which we are partnering with the Smithsonian Institute and the California Community Colleges System, and the Next Generation Courseware network, in which we are partnering with Smart Sparrow.

In active partnership with edX, with me as Director of DANCE: Discussion Affordances for Natural Collaborative Exchange, I have worked to build community around dissemination of research and resources enabling large scale deployment of discussion based learning practices at scale, including in Massive Open Online Courses (MOOCs). Over 5,600 individuals have participated in the growing DANCE community in some way since the launch of the community website in Spring 2015, and hundreds of return visitors participate in events (such as the monthly online talk series) or access software or publication resources on the community website each month. The DANCE community provides a “go to” place for resources developed by my own lab and other collaborating organizations. My own lab’s research in the area of supported collaborative learning in MOOCs includes analyses of data from dozens of MOOCs as well as completed deployment studies in six different MOOCs. An example is a recent collaborative effort with the Smithsonian Institute where one of my PhD students has developed a team-based learning component within a Smithsonian course as a test of interventions fine-tuned and rigorously validated first in Amazon’s Mechanical Turk. We are actively preparing for additional deployments including two more with the Smithsonian Institute and one in partnership with George Siemens at the University of Texas at Arlington and the Whitehouse’s ConnectedEd Initiative. Our deployed conversational agent facilitated collaborative chat intervention has been demonstrated to significantly reduce attrition (sometimes reducing probability of dropout at the next time point after experience of collaborative interaction by more than a factor of two). Attrition is noted to be one of the major challenges of MOOC-based instruction.

Carolyn Penstein Rose (cprose@cs.cmu.edu)/ Carnegie Mellon University