Replication Data for The Moral Check: A Systematic Review and Strategic AI Governance for the Pacing Problem
Description
This dataset provides the complete open-access replication archive for the PRISMA 2020 systematic literature review: The Moral Check: A Systematic Review and Strategic AI Governance for the Pacing Problem. Dataset Contents: 1. Master_518_FullText_Appraisal_Matrix.xlsx: Complete multi-tab appraisal dataset of all 518 candidate reports retrieved, documenting live DOIs, reviewer blind screening votes, and granular exclusion justifications across the 4 PRISMA rationales (142 normative, 118 conceptual, 86 whitepapers, 42 post-hoc). 2. MMAT_2018_Appraisal_Matrix_130_Studies.xlsx: Detailed study-by-study methodological quality evaluations across all five core categories of the Mixed Methods Appraisal Tool (MMAT 2018) for all 130 qualified empirical primary studies (>= 80% quality score). 3. Search_Queries_All_5_Databases.xlsx: Exact Boolean query strings, proximity operators, date filters (2017-2026), and hit yields across Scopus (1,842), Web of Science (1,124), IEEE Xplore (618), ACM Digital Library (486), and PubMed/MEDLINE (216), yielding 4,286 raw bibliographic records. 4. The_Moral_Check_179_Final_References.ris: Citable RIS reference library for all 179 references cited in the manuscript with active DOIs. 5. The_Moral_Check_518_FullText_Screened.ris: Complete 518 full-text retrieved corpus tagged with screening decisions (Included vs Excluded). 6. SAGE-X_Ontology.owl: Formal OWL 2 DL ontology formalizing the Three-Tier Architecture, behavioral actor veto keys, and Moral Check Index (MCI) admission controllers. Permanent OSF Archive: https://osf.io/p69vt Stanford WebProtégé Project: https://webprotege.stanford.edu/#projects/1d767dc6-998c-4460-b0a4-9c69ed76eea0
Files
Steps to reproduce
1. Multi-repository search executed across Scopus, Web of Science, IEEE Xplore, ACM Digital Library, and PubMed/MEDLINE on January 31, 2026, using custom Boolean strings (Search_Queries_All_5_Databases.xlsx) yielding 4,286 raw records. 2. Automated deduplication conducted via Rayyan and Mendeley, removing 1,372 duplicates to establish 2,914 unique records (Deduplication_Audit_Log.xlsx). 3. Double-blind title and abstract screening conducted independently by two reviewers in Rayyan, excluding 2,396 records (Cohen's kappa = 0.88) and retrieving 518 candidate full-text reports. 4. Independent full-text appraisal of all 518 retrieved studies (Master_518_FullText_Appraisal_Matrix.xlsx), excluding 388 studies across 4 explicit PRISMA rationales (142 normative discourse, 118 conceptual blueprints, 86 vendor whitepapers, 42 post-hoc remedies; Cohen's kappa = 0.91). 5. Methodological quality evaluation of all 130 qualified primary empirical studies using the Mixed Methods Appraisal Tool (MMAT 2018; MMAT_2018_Appraisal_Matrix_130_Studies.xlsx), verifying that all included studies satisfied >= 80% quality score. 6. Semantic formalization of the SAGE-X Three-Tier Architecture into OWL 2 DL ontology (SAGE-X_Ontology.owl), valida ted via Stanford WebProtégé and Pellet reasoner.
Institutions
- INTI International UniversityNegeri Sembilan, Kampung Baharu Nilai