Anonymous data and analysis package for classroom GenAI mediation in EFL revision learning

This public dataset accompanies a double-anonymous manuscript submitted to Computer Assisted Language Learning. Depositor and creator information are intentionally anonymized. The package contains de-identified analysis data, processed coding and feedback records, metadata, validation outputs, analysis outputs, and scripts for reproducing the manuscript-level analyses. Raw essay texts, raw feedback texts, raw interview transcripts, identity mappings, and author-identifying submission materials are excluded.