Coding reliability dataset for: Representation-to-AI Transformation in K–12 Generative AI Learning: A Theory-Building Systematic Review of Semantic Transformation Mechanisms

This dataset provides the complete double-coding matrix, PRISMA 2020 checklist, and search strategy supporting the systematic review "Representation-to-AI Transformation in K–12 Generative AI Learning: A Theory-Building Systematic Review of Semantic Transformation Mechanisms." It includes: (1) study-level tier classification (Core/Supporting/Context) for two independent coders and consensus tier for all 18 included studies; (2) the full semantic transformation unit (STU) coding matrix (18 studies x 10 STUs = 180 cells) with pre-consensus and consensus scores; (3)evidence-weighting consensus scores; (4) inter-rater reliability statistics (Cohen's kappa); (5) the completed PRISMA 2020 checklist; and (6) the full database-specific Boolean search strategy.