MedGaugeVersions
Changelog
Newest first. Results under different benchmark versions are not comparable. How each object changes is set out on Updates; this list is also published as an Atom feed.
| Date | Type | Change |
|---|---|---|
| 2026-10-29 | Methodology | First publication. Methodology 0.1.0: the check as the unit of measurement; ten dimensions (M1–M10) reported each on its own, with no total; ten task labels (T1–T10); three runs per task, a run error counted as a failed run with its attribution, and a fault on the initiative's side voiding that task version for every subject; the exam conclusion, stated with its exact lower bound 0.05^(1/n); Clopper–Pearson intervals for 0/1 task figures and task-level BCa bootstrap for task means; paired, Holm-adjusted comparisons with a compact letter display. No level threshold is set, so no level is issued. Charter 0.1.0. Benchmark version 0.1.0 of Clinical Temporal Audit, Bioinformatics QC and Guideline Evidence Check, frozen under [email protected]; Medical Exam Item Review is withheld while its answer key is under review. Registries of capabilities, harnesses, tools, standards, benchmarks and grading references, with schemas at version 0.1.0. Contributions by e-mail, with the task-submission form, its field table, an example and a submission check; an Updates page; this changelog as an Atom feed. |