XAI Evaluation Metrics in Explainable Healthcare AI — Master Dataset

Complete evidence base of the systematic literature review conducted for the MSc thesis "An Exploratory Study of Evaluation Metrics in Explainable Healthcare AI" (Vrije Universiteit Amsterdam, MSc Information Sciences, 2026), following the PRISMA 2020 guidelines. The workbook contains four tabs: (1) README — reader's guide; (2) Papers — the 36 included papers (2021–2026) with identifiers and bibliographic details; (3) Raw records — 524 metric-use records, one row per metric use in a paper, including original and standardized metric names, definitions as stated, metric type (Explicit/Implicit), original and validated XAI paradigm, paradigm confidence (High/Medium/Low), explanation technique, evaluation method, stakeholder, and measurement character; (4) Canonical metrics — the 257 consolidated metrics after name standardisation with type, paper counts, dominant paradigm, and operational definitions. The stored paradigm label "Intrinsic" corresponds to "Pre-hoc" in the thesis text. Corpus note Paper IDs run P001–P038 and are not contiguous: P030 and P033 are absentbecause those records were excluded during screening and the remaining IDs werenot renumbered. The corpus therefore holds 36 papers. All but one carry a DOI.