NSF Award
2415032
As understanding of the world becomes ever more complex, it is important that people understand how and why phenomena occur, rather than simply knowing that they occur. For example, understanding why a large unvaccinated population is more likely to give rise to mutations in viral infections, or why polyfluorinated alkyl substances (PFAS) are so difficult to remove from drinking water, are the first steps in addressing such problems. Thus, it is important to support undergraduate students in learning to construct mechanistic explanations and models to explain how and why things happen. Students who engage in these activities are more likely to learn deeply and be able to use their knowledge in new situations. Thus, helping students engage in mechanistic reasoning will not only support students as they work to connect relevant ideas, but will also provide evidence about the depth and interconnectivity of student understanding. This project aims to serve the national interest by determining how to support large numbers of undergraduate chemistry students as they construct mechanistic explanations and models, particularly in large introductory courses that often act as a barrier to Science, Technology, Engineering, and Mathematics (STEM) success. To achieve this goal, the project team will design tasks that have the potential to elicit student explanations about how and why phenomena occur. Students’ responses to such tasks are typically time consuming for instructors to grade, and particularly as the phenomena become more complex, it may become difficult for students to connect all the ideas involved. To provide support for these complex tasks, the project team will train and implement generative artificial intelligence (AI) chatbots to provide feedback to students. It is expected that these chatbots will be able to accurately characterize students’ responses and provide feedback in a variety of ways, including in the form of a Socratic dialogue.<br/><br/>Evidence suggests that having undergraduate students construct mechanistic explanations in the context of formative tasks is an equitable approach to instruction for courses that typically function as a gateway to STEM careers. To achieve equitable instruction, it is important to support students as they regularly engage with tasks designed to extend and connect their knowledge. Thus, the project team from Michigan State University (MSU) will implement a modified evidence-centered design process to determine both the cognitive and epistemic resources students need to construct mechanistic explanations and to articulate the evidence of understanding expected to be elicited for a range of complex tasks. Using those design specifications, generative AI feedback systems will be designed to support students’ knowledge construction and knowledge use. The AI chatbots will be implemented in general chemistry courses at MSU and the project team will characterize the ways in which students’ learning and perceptions of learning are impacted by interacting with the chatbots. These findings will be used to develop a framework for the design of explanation tasks involving complex phenomena and the accompanying AI feedback systems to best support students. Results of the investigations have the potential to transform STEM courses by advancing knowledge of how to design and engineer instructional tasks that both provide opportunities to connect and use knowledge and provide richer insights into what students know and can do. The NSF IUSE: EDU Program supports research and development projects to improve the effectiveness of STEM education for all students. Through its Engaged Student Learning track, the program supports the creation, exploration, and implementation of promising practices and tools.<br/><br/>This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
