Deidentified data and analysis code for "LLM-assisted assessment item writing in medical education: a randomized pilot trial
Abstract
This repository contains analysis code underlying the psychometric analyses of a randomized pilot study comparing LLM-assisted and conventional medical assessment item writing. It includes item-level psychometric outcomes, faculty-level aggregated outcomes used for exact permutation testing, and the corresponding R analysis code. Examination stems, response options, answer keys, and other confidential examination materials are not included.
// Source
Authors: Vincent Hennion, David Attali, Damien Roux, Henri Duboc, Jennifer Truchot, Alexandre Hudon, A. Moussa, Bruno Étain, Cédric Lemogne, Charles Ouazana Vedrines
Institutions: Université de Montréal, Inserm, Université Paris Cité, Sorbonne Paris Cité, Assistance Publique – Hôpitaux de Paris, Institut universitaire en santé mentale de Montréal