Connection making between multiple graphical representations: A multi-methods approach for domain-specific grounding of an intelligent tutoring system for chemistry

Abstract

Making connections between graphical representations is integral to learning in science, technology, engineering, and mathematical (STEM) fields. However, students often fail to make these connections spontaneously. Intelligent tutoring systems (ITSs) are suitable educational technologies to support connection making. Yet, when designing an ITS for connection making, we need to investigate what concepts and learning processes play a role within the specific domain. We describe a multi-methods approach for grounding ITS design in the specific requirements of the target domain. Specifically, we applied this approach to an ITS for connection making in chemistry. We used a theoretical framework that describes potential target learning processes and conducted a series of four empirical studies to investigate what role graphical representations play in chemistry knowledge and to investigate which learning processes related to connection making play a role in students' learning about chemistry. These studies combined multiple methods, including knowledge testing, eye tracking, interviews, and log data analysis. We illustrate how our findings inform the design of an ITS for chemistry: Chem Tutor. Results from two pilot studies done in the lab and in the field with altogether 99 undergraduates suggest that Chem Tutor leads to significant and large learning gains on chemistry knowledge.