This is the document you point to when someone challenges a verdict. The methodology is not partisan. It is applied consistently regardless of who made the claim. The standards existed before any claim was graded and will apply to any claim graded in the future.
Most fact-check organizations use binary or three-way systems that collapse important distinctions. DECLASSIFIED uses four verdicts because four verdicts are necessary to capture the actual range of how claims can be wrong. FALSE and MISLEADING are not the same thing. INSUFFICIENT INFO is not a failure of analysis — it is the most intellectually honest verdict available when the evidentiary record is genuinely incomplete.
Every claim and every player citation is evaluated against a four-tier scale based on the source's proximity to the primary record. The scale is not about prestige — a peer-reviewed journal can be WEAK if its methodology is contested; a government document can be STRONG even if you dislike the government.
| Tier | Definition | Examples |
|---|---|---|
| STRONG | Primary source, official record, or direct documentation. The evidence itself, not a report about the evidence. | Congressional testimony, court filings, official transcripts, medical examiner reports, authenticated recordings, NARA documents |
| CAREFUL | Expert consensus, major institutional reporting with multiple sources, or peer-reviewed analysis with established methodology. | Bipartisan congressional committee reports, Federal Reserve analyses, BLS data releases, multi-outlet verified reporting |
| WEAK | Contested, partisan, or secondary-source basis only. May be accurate but cannot be independently verified against a primary record. | Single-outlet reporting without sourcing, op-eds, advocacy organization analyses, media characterizations of sealed proceedings |
| INSUFFICIENT | No credible or identifiable evidentiary basis. The claim is unsourced or its source is not traceable. | "People are saying," anonymous social media posts, conspiracy sites, claims whose primary source cannot be identified |
DECLASSIFIED grades claims against primary sources wherever they exist. Primary sources include: official government documents, authenticated recordings, court filings, sworn testimony, medical examiner records, NARA materials, and official statistical releases from agencies like BLS, CBO, and the Federal Reserve.
When a primary source is unavailable, the highest available secondary source is used and the citation tier is adjusted accordingly. The absence of a primary source does not automatically result in INSUFFICIENT INFO — it results in a lower citation weight applied to the verdict.
Why is "unprecedented persecution" MISLEADING and not FALSE? Because the factual predicate — that no former president had been federally indicted before 2023 — is accurate. MISLEADING captures the accurate predicate while flagging the false inference drawn from it. The verdict is more precise than FALSE would be.
Why is the presidential declassification by thought argument rated FALSE? Because Executive Order 13526 requires a formal written process, no court has recognized declassification by mental intent, and the recovered documents bore active classification markings at the time of seizure. The claim is directly contradicted by three independent primary-source records.
Why don't we grade claims in both directions — why not fact-check Democrats? We do, and we will. DECLASSIFIED by Sean is currently focused on a specific body of documented claims from a specific period. The methodology applies to any claim from any speaker. The framework is not partisan. The dataset is currently focused.