Verification Judgment in AI-Mediated Higher Education: Bounded Construct Landscape Scan Materials

Supporting materials for a bounded construct landscape scan reported in the article “Verification Judgment in AI-Mediated Higher Education Learning Environments: A Conceptual Construct” (Yan, 2026, Journal of Computing in Higher Education, under review).

The scan covered five bibliographic databases (OpenAlex, Crossref, Semantic Scholar, arXiv, DBLP) for English-language records published between 30 November 2022 and 3 May 2026. Identification yielded 312,828 records; deduplication retained 285,783 unique records; title-and-abstract eligibility screening retained 17,638; focality filtering retained 15,045; construct charting against seven focal constructs (academic integrity, critical thinking, AI literacy, verification, calibration, evaluative judgement, sycophancy) produced a final corpus of 5,306 records.

Contents:
– SupplementaryFile1_ScanProtocol.docx: Full protocol, codebooks, search strings, schema-agreement analysis, saturation analysis, co-occurrence matrix.
– SupplementaryFile2_Corpus.csv: Machine-readable corpus of 5,306 records with charting variables under both Pass A (strict) and Pass B (extended) schemas.
– SupplementaryFile3_PipelineCode.tar.gz: Seven-stage Python pipeline source code for independent replication, plus README.

See README.md for detailed contents description.