AI-Adapted Grading

Recruiting Instructors

Overview

Generative AI can now complete most economics homework, which raises the question of how much weight work done outside the classroom should carry in a course grade. This multi-institution field experiment tests whether shifting grade weight from homework to in-class assessments improves student learning and reduces stress, while holding the AI policy and the way assignments are graded constant. Each participating instructor teaches the same economics course in two of three terms (Spring 2027, Fall 2027, and Spring 2028), once with their usual weights and once with the weight on outside-of-class work cut in half, in a randomly assigned order. Learning is measured by comparing student performance on each instructor’s own final exam, which stays the same across their two terms. Students also complete short surveys on their AI use, stress, and perceptions of the course. We aim to recruit about 45 instructors and welcome US-based economics instructor who expects to teach the same course in two of the three terms.

For full details, read the project’s I2D proposal.

Interested in participating in this project? Fill out this survey now!