PIs: Emma Mercier, Luc Paquette
University of Illinois at Urbana-Champaign
This project is focused on the teaching of collaborative problem solving activities in introductory engineering courses and builds on a prior project to design tools for collaborative sketching in these courses. The project is based on a recognition of the importance of collaborating in engineering, the need for student to learn this skill, the value of collaborative learning tasks for engaging students in authentic problem solving activities, and the difficulty that graduate student teaching assistants (TAs) encounter when trying to teach in this way. There are two parts to the technology innovation. The first part is a set of tools for the teaching assistants, to help them manage the classroom technologies, and to help them understand how to intervene in groups who are struggling with the content or collaborative processes. The second part is a set of tools for the students. Building on the collaborative sketch software previously developed, prompts to support their collaborative processes will be embedded in the software students will use, based on analysis of the logfiles that help determine who needs what prompts when. Research goals include understanding how receiving prompts changes the nature of students’ collaborative activity, and how receiving insight into the difficulties students are having helps TAs learn about to foster collaborative learning in their classes.
The PIs are addressing the difficulties encountered implementing collaborative learning activities in engineering courses by designing and studying tools for TAs and students in these classes. Through an iterative design approach, the PIs will design and study tools for TAs to orchestrate the classroom and collaboration activities and to tools for students which support their collaborative problem solving processes. The PIs will investigate the use of learning analytics in evaluating the collaborative practices of students using these tools; in particular, logfiles will be examined for collaborative indicators based on prior research on collaborative processes, then clustered to look for patterns of engagement, and finally used to create regression models of successful collaboration processes using machine learning techniques. Cross-validation of the models will be done with both logfile and video data to avoid overfitting. These insights will be provided to TAs to examine whether such information is helpful in determining how and when to intervene in groups. Findings from the research will provide insight into: 1) The knowledge that TAs need in order to successfully implement collaborative problem solving in undergraduate courses; 2) Whether TAs can learn more about collaborative problem solving with the support of tools aimed at helping them implement this form of pedagogy; 3) Whether students can learn collaborative problem solving skills through embedded prompts during multi-week collaborative activities and 4) The potential of analytics in determining when and how to reduce the collaboration supports from groups.