Generative AI to Analyze Students' Scientific Reasoning Tasks

The project aims to investigate the potential of Large Language Model in scoring students’ constructed responses for scientific reasoning assessment tasks. Our research team explores strategies for developing algorithms using Large Language Models (e.g., GPT) to automatically score students’ constructed responses on assessments that target three-dimensional (3D) science learning. These assessments require students to integrate scientific practices, crosscutting concepts, and disciplinary core ideas to make sense of phenomena or solve complex problems.