EEF/NFER Randomised Trial: ChatGPT Cuts Science Teachers' Lesson-Prep Time by 31%

The Education Endowment Foundation (EEF) and the Hg Foundation commissioned the National Foundation for Educational Research (NFER) to run a school-randomised controlled trial testing whether ChatGPT could reduce teacher workload. During the 2024 summer term, 259 Year 7 and Year 8 science teachers across 68 English secondary schools were randomly assigned either to use ChatGPT (with an online guide on effective use) for lesson and resource preparation, or to prepare lessons as usual, over a 10-week period.

Teachers in the ChatGPT group reported spending an average of 56.2 minutes per week on lesson and resource planning, versus 81.5 minutes per week in the comparison group — a saving of 25.3 minutes per week, or 31% of planning time (69% of the comparison group's time). An expert panel reviewed a sample of the resulting lesson resources without knowing which group had produced them, and found no evidence that quality differed between the ChatGPT and non-ChatGPT groups. Teachers most commonly used ChatGPT to generate questions or quizzes and to find new activity ideas, rather than for whole-lesson planning. NFER published the evaluation report on 12 December 2024.

The trial has an important scope limitation the authors flag directly: teachers only applied their assigned approach to part of their timetable, not their full teaching load, so the time savings may not hold at full-scale, sustained use. The study also measured teacher-reported time and resource quality, not downstream student learning outcomes, and covered a single subject (science) at one stage of English secondary education.

Read the full analysis: https://www.nfer.ac.uk/publications/chatgpt-in-lesson-preparation-a-teacher-choices-trial/