More Confident, But More Competent? Self-Efficacy Calibration in AI-Assisted Writing This project hosts the de-identified data, and analysis code for a randomized experiment examining how generative A

More Confident, But More Competent? Self-Efficacy Calibration in AI-Assisted Writing This project hosts the de-identified data, and analysis code for a randomized experiment examining how generative AI–supported writing changes the calibration between university students’ writing self-efficacy and writing performance. Participants (N = 270) were assigned to no-AI, free-AI,… EdArXiv (education prep