NLP Methods May Actually Be Better Than Professors at Estimating Question Difficulty

TL;DR

This study demonstrates that Large Language Models, especially when using uncertainty-based supervised learning, can outperform professors in estimating exam question difficulty, leading to improved assessment quality.

Contribution

It introduces a novel approach using LLM uncertainties and supervised learning to better estimate question difficulty compared to human judgment.

Findings

LLMs outperform professors in difficulty estimation.

Uncertainty-based supervised learning improves accuracy with minimal data.

Professors have limited ability to distinguish question difficulty.

Abstract

Estimating the difficulty of exam questions is essential for developing good exams, but professors are not always good at this task. We compare various Large Language Model-based methods with three professors in their ability to estimate what percentage of students will give correct answers on True/False exam questions in the areas of Neural Networks and Machine Learning. Our results show that the professors have limited ability to distinguish between easy and difficult questions and that they are outperformed by directly asking Gemini 2.5 to solve this task. Yet, we obtained even better results using uncertainties of the LLMs solving the questions in a supervised learning setting, using only 42 training samples. We conclude that supervised learning using LLM uncertainty can help professors better estimate the difficulty of exam questions, improving the quality of assessment.