Skip to main content
ASP2024: A Virtual Conference Main banner
Final Program


Schedule-at-a-Glance
Icon Legend
This session is not in your schedule.
This session is in your schedule. Click again to remove it.
Back
Session VIDEO link

Teaching Astronomy – New Opportunities, Challenges, Tools, and Innovations

Session: 12C: 10-Minute Orals (4)

12C-1 - Using Large Language Models to Assess Student Writing Assignments

Saturday, August 24, 2024
10:50 AM – 11:40 AM US PDT
Abstract: Although writing assignments are an excellent way to promote and assess student learning, it is difficult to implement student writing in large classes, and nearly impossible to provide iterative feedback. For large online classes, such as Massive Open Online Classes (MOOCs) peer grading has been the only solution. Unfortunately, peer grading can be unreliable and often peer graders do not leave any useful content-specific feedback. In this project we used Large Language Models (LLMs) GPT-3.5 and GPT-4 to test the use of Machine Learning as an alternative to peer grading and feedback. The results show that LLMs reliably provide comprehensive and accurate feedback similar to feedback from instructors. Our findings also show that LLMs assigned grades that were consistent with instructor scores, and more accurate and reliable than peer graders.