Transforming GenAI Policy to Prompting Instruction: An RCT of Scalable Prompting Interventions in a CS1 Course
A semester-long RCT found structured prompting instruction improved skills and predicted higher final exam scores.
Deep Dive
Researchers from Carnegie Mellon University conducted a randomized controlled trial (RCT) with 979 CS1 students. They tested four instructional conditions based on the ICAP framework to teach 'prompting literacy'—using GenAI as a learning tutor. Results showed all conditions improved prompting skills, and gains in immediate post-tests predicted higher final exam scores. The study provides scalable methods to transform classroom AI policies into effective learning instruction.
Why It Matters
Offers educators a proven, scalable framework to turn AI from a crutch into a legitimate teaching assistant that improves outcomes.