Across fifty years and 49 yearly India entries, Amnesty's method never changes: there is no codebook, no way to check individual claims inside the entry, and an evidence base that is entirely Amnesty's own reporting. The events get heavier — the 1975 Emergency, Punjab, Kashmir, and the recent wave of demolitions and minority-targeting — but the grade does not move. Strong organizational independence and open finances keep it out of the basement; the thin, self-referential sourcing keeps it Deficient. The grade describes the document a reader actually holds — one country section in a global yearbook — not Amnesty's whole institution.
Composite score 4.6 / 10 — Deficient. Raw weighted score was 4.6.
| Dim | Dimension | Score | Weight | Flag |
|---|---|---|---|---|
| D1 | Definitional Precision | 4 | 17.9% | No published codebook or operational definitions in any of the 49 entries. |
| D2 | Classification Rigor | N/A | 0% | — |
| D3 | Case Capture & Sampling | N/A | 0% | — |
| D4 | Coverage Symmetry | 6 | 22.4% | — |
| D5 | Source Independence | 6 | 14.9% | — |
| D6 | Verification Standards | 3 | 26.9% | The only score that moves across the series: 3 (Era A) → 4 (Era B) → 3 (Era C). It is also the heaviest dimension at 26.9%. |
| D7 | Transparency & Governance | 7 | 7.5% | — |
| D8 | Counter-Evidence | 3 | 10.4% | — |
| Composite Score | 4.6 | Deficient | ||
Scope: Across most of the series, factual claims — numbers detained, killed, or tortured — carry no in-text sourcing. The only verification trail that ever appears (Era B, 2000–2013) points back to Amnesty's own reports.
Scope: Across 49 India entries and five decades, the dimensions that measure method — D1, D4, D5, D7, D8 — do not change. Only D6 moves, and only by one point. Score variation tracks events and formatting, not methodological reform.
Scope: No codebook, operational definitions, or borderline-case guidance appears in any of the 49 entries. A reader cannot reconstruct the rule that decided which cases were included.
Scope: Zero external citations at the entry level throughout the series. The evidence base is Amnesty's own field research, reported by Amnesty, with no independent check built into the document.
Scope: Era C (2014/15–2024/25) is scored as presented in the available entry text. Modern entries carry endnote sourcing not captured in the reviewed slice; if captured, Era C verification would likely rise one point to match Era B. Flagged rather than silently resolved.
No published codebook or operational definitions in any of the 49 entries.
Amnesty applies institutionally-defined terms of art — 'prisoner of conscience,' 'extrajudicial execution,' 'arbitrary detention,' 'torture' — that carry established Amnesty meanings. But the yearbook entry republishes no codebook, no operational definitions, and no borderline-case guidance: the document review finds a definitions or glossary section in 0 of 49 entries. Terms are applied editorially inside the entry, so a reader cannot reconstruct the inclusion rule that placed a given case in the report. This is the band-4 reading — terms used without in-document operational specificity, relying on an implicit 'we know it when we see it' standard. The score sits two points below the Freedom House India chapter (6), which is backed by a published 25-question framework and a companion methodology document; Amnesty International's entry has no such companion document.
Not applicable for TYPE 6 Advocacy Documents. The excluded weight is redistributed proportionally across D1, D4, D5, D6, D7, and D8.
Not applicable for TYPE 6 Advocacy Documents. The excluded weight is redistributed proportionally across D1, D4, D5, D6, D7, and D8.
The parent document, 'The State of the World's Human Rights,' walks every country A–Z, so cross-national coverage is broad rather than particularist. Within India, Amnesty's evaluative criteria — torture, arbitrary detention, unlawful killing, unfair trial — are identity-neutral and applied to both state actors and armed or non-state groups; entries across the series carry explicit 'Abuses by armed groups' sections. The criteria pass the Swap Test: removing identity markers does not change how they function. What shifts over time is editorial emphasis, not the criteria — length-normalized mentions of Muslims rise from 0.92 (Era A) to 2.79 (Era C) per 1,000 words, and Dalits from 0.05 to 2.07, tracking events on the ground. Coverage distribution is not benchmarked against base-rate data, and recent entries lean heavily on a single victim-group framing, which holds the score below the 7–9 band.
Amnesty is institutionally independent of the Indian government and Indian political parties and famously refuses government funding for its research — a strong structural-independence signal — and no circular advocacy loop with other NGOs is visible at the entry level. The evidence base, however, is Amnesty's own field documentation, and at the entry level there are essentially no external citations (URLs = 0 throughout; the Era-B 'AI country reports' references are self-citations to Amnesty's own fuller reports). Under the rubric's source-type rule, a primary-research institution citing its own documentation counts as internal documentation rather than circular sourcing, which keeps this out of the 4–6 floor. Score of 6 reflects independence from the subject combined with an entirely self-referential evidence base.
The only score that moves across the series: 3 (Era A) → 4 (Era B) → 3 (Era C). It is also the heaviest dimension at 26.9%.
Under the adapted TYPE 6 criterion, citation accuracy replaces dataset replication: statistical claims are checked against their cited sources. The problem is that there are almost no citations to check. In Era A (1973/74–1999, scored 3) factual assertions — numbers detained, killed, tortured — carry zero in-text sourcing; claims rest entirely on institutional trust. In Era B (2000–2013, scored 4) each entry closes with an 'AI country reports and visits' list carrying AI-Index numbers — a consistently published, documented pathway to Amnesty's own fuller reports; still internal, but a real reference trail that lifts the entry to the bottom of band 4. In Era C (2014/15–2024/25, scored 3) the in-text reference lists disappear and the slice reverts to unsourced assertion. The series-level score of 3 reflects the dominant Era A and Era C state. Note: Era C is scored as presented in the available slice text; modern Amnesty entries carry endnote sourcing that the extraction did not capture, and if captured Era C would likely rise to 4. Under the D6 < 7 rule, Research-Grade is blocked for the entire series — moot here, since no era approaches 8.0.
Amnesty is a large membership organization that publishes audited financial statements and, critically, declines government and corporate funding for its research, funding itself through members and individual donors. Governance is a genuine board structure, not a founder-controlled operation. The entry itself carries no per-report funder disclosure and no in-document data-ethics policy, which holds it below the top band. Score of 7 reflects strong organizational transparency with no entry-level disclosure apparatus.
Adapted for TYPE 6, limitations and corrections are weighted above counter-evidence engagement. No entry across the series contains a limitations section, a corrections policy, or engagement with perspectives that complicate Amnesty's assessment. The entries do record positive developments where they occur ('in a welcome step, the Supreme Court suspended…'), which keeps the score off the floor — Amnesty is not impervious or conspiratorial about criticism; it simply does no methodological self-examination inside the document. Band 1–3.