MAAC Study 3 Psychometric Validation Dataset: Instrument v4.7 Baseline and Revision Cohorts

This dataset contains the source data supporting the psychometric validation reported in Study 3 of the Multi-Dimensional Assessment for AI Cognition (MAAC) dissertation series. The primary dataset is the full cohort psychometric validation run (n = 2,196) used to assess the internal structure of the nine six-item MAAC dimensions across hypotheses H1–H5 and research questions RQ1–RQ5. Additional files document the iterative instrument revision program leading to the accepted v4.7 endpoint, including targeted revision cohorts for Tool Execution, Knowledge Transfer, and Hallucination Control Q6 redesign. The final accepted baseline state is 53 of 54 items at factor loading ≥ 0.40 and 8 of 9 dimensions at acceptable or good CFA fit. The accepted manuscript endpoint is instrument v4.7. This dataset supports replication of findings reported in: "Construct Validity of AI-Administered Cognitive Assessment: Psychometric Validation of a Nine-Dimension Process-Oriented Evaluation Instrument."