EXP: Students Authoring Intelligent Tutoring Systems for Constructing Models of Ill-Defined Dynamic Systems

May 21, 2013

PIs: Kurt VanLehn, Daniel Childers
Arizona State University

The PIs, one expert in intelligent tutoring systems, and the other an expert on ecosystems, are investigating the ways model construction can aid in the understanding of complex dynamic systems, specifically ecosystems and systems involved in sustaining life on our planet. The PIs are particularly interested in understanding the affordances of model construction in promoting such learning and the challenges in doing so, and they are doing that by having undergraduates construct and explore models related to issues in sustainability during an introductory sustainability course. Model construction and exploration are done in the context of an intelligent tutoring system (called Lait). Learners switch between constructing models themselves and using the models of others as learning aids. Model constructors know they are constructing their models for others to learn from, and as others are using those models to learn, they may interact (online) with the model constructors to discuss the reasoning behind designing the models. The tutoring system provides scaffolding for model building and for model exploration, supports collaboration around modeling issues and issues related to particular models, and runs the models that are created. The goal is to promote understanding of the systems being modeled, skill at model construction, and understanding of the epistemology of modeling. The first year of research is taking place in the lab; research in later years is taking place in classrooms, specifically in the introductory interdisciplinary sustainability course, which enrolls 300 students per semester.

Systems thinking is notoriously difficult and notoriously difficult to learn. Thinking about sustainability is equally complex, and the target domain here is sustainability. There is potential for this work to shed light on how to better help people understand and think about systems and about interactions between and interconnectedness of the systems that sustain our lives on earth. In addition, the literature on learning suggests that learners should learn more from constructing models and making them work than from simply running the models of others as simulations. But up to now, technology has not been up to making modeling by learners accessible. Most research in the area of simulation and modeling to promote learning has therefore focused on the roles of simulation in promoting learning. This project is exploring ways of making systems modeling more accessible to those who are not already competent programmers and will uncover some of the affordances of modeling for promoting learning and the circumstances under which modeling can promote learning about particular systems and learning to be a systems thinker.