Downloads
The registry is published as a canonical JSON dataset, a grade-matrix CSV, and a printable PDF.
- instrument-evidence-base-v0.9.0.json: the canonical dataset (31 records at schema v0.7: full findings, grades with evidence state, rubric version and review flags, provenance, licence classes with archived steward pages, citations, corrections, errata, right of reply).
- instrument-evidence-matrix-full.csv: the 27-instrument grade matrix, generated from the dataset at each build, with the criterion split and each cell's indirectness flag or absence type.
- RUBRIC-v1.6.md: the grading rubric every cell cites.
- ADMISSION-v1.1.md: the admission rule, how an instrument gets a record at all.
- instrument-registry.pdf: a printable export of the matrix and record summaries.
Relations between records (item-of, short-form-of, screens-for, corresponds-with, embeds) are in the dataset as relations on each
record, forward only, each with the study it rests on; the single-item measures page renders them.
Previous versions
Earlier dataset files stay at their URLs so that citations resolve: instrument-evidence-base-v0.8.0.json, instrument-evidence-base-v0.7.0.json, instrument-evidence-base-v0.6.0.json, instrument-evidence-base-v0.5.0.json, instrument-evidence-base-v0.4.0.json, instrument-evidence-base-v0.3.0.json, instrument-evidence-base-v0.2.json, instrument-evidence-base-v0.2.2.json, instrument-evidence-base-v0.2.1.json, instrument-evidence-base-v0.1.json. What changed between versions is on the corrections page under dataset changes.
How to cite
Licensing
Registry text is licensed CC BY 4.0; the dataset structure is licensed Apache 2.0. Instrument item text is never reproduced in this registry; each instrument's own licence status is recorded on its page, verified against the steward's current terms where possible.