Faculty Sponsor: Dr. Antonio Laverghetta
Gabriel Diaz
Gabriel is a second-semester senior (’26) majoring in Computer Science , with a minor in Data Analysis. He is originally from Eastern Pennsylvania, and has plans to move to New England after he is finished at Wesleyan. During his time here, he took part in Aikido and worked with the Events and Conferences team in Usdan. Gabriel wants to go into Data Engineering in the future, either within Fintech or the Medical Field.
Abstract: Recent advances in large language models (LLMs) have created new opportunities for automating the generation of educational assessment questions. However, relatively little is known about how prompting strategies and retrieval-augmented generation (RAG) influence the quality of generated questions, or whether LLMs can reliably evaluate question quality themselves. This project investigates the content validity of introductory psychology multiple-choice questions generated under multiple prompting conditions and compares automated LLM evaluations with expert human judgments. The results provide insight into the feasibility of using LLMs to support both assessment generation and quality evaluation in educational settings.
gdiaz_project26