Observed content or event
Does the post, document, outage, recording, or event exist?
Twelve reviewed research packets
Separate the event,
the claim, and the attribution.
A publication-safe method for separating observed content, claim status, coordination, actor identity, sponsorship, intent, reach, effect, source reliability, analytic confidence, oversight capability, remedy, leaderless-network roles, metadata inference, privacy, and correction.
Direct answer
What is the IARPG evidence-attribution method?
It is a non-operational editorial method that prevents one observation from becoming an unsupported claim about coordination, identity, sponsorship, intent, reach, or effect. It records claim status, analytic confidence, and source reliability separately; counts independent evidence origins rather than repeated headlines; compares oversight by actual powers and remedies; and preserves corrections without scoring people, communities, organizations, or countries.
- Primary rule
- Stop at the highest supported attribution rung
- Evidence axes
- Claim status · analytic confidence · source reliability
- Source custody
- 12 exact reports with one-to-one reviewed wrappers
- Publication state
- Reviewed with limitations · no live attribution service
How to use this page
Use the page as a claim-checking sequence.
Begin with the smallest observable proposition. Trace its source chain, identify the highest supported attribution rung, record the three evidence axes, inspect rights and remedy, and preserve later corrections as new records.
-
01
Bound the propositionWrite one claim with a date, actor, action, object, and scope.
-
02
Trace the origin chainFind the earliest accessible record and identify repeated derivatives.
-
03
Stop at one rungDo not let content, coordination, identity, sponsorship, intent, reach, or effect collapse together.
-
04
Record limits and remedyExpose counterevidence, privacy risks, correction routes, oversight, and unresolved questions.
Eight-rung attribution ladder
Each rung needs
its own evidence.
An observed post or outage can be real while its claimed actor, sponsor, intent, scale, or effect remains unresolved. Public wording must stop at the highest rung supported by direct, independent, and attributable evidence.
Content status
Is a discrete factual claim supported, contradicted, misleading, or unresolved?
Coordination
Did multiple assets act together under shared control or planning?
Actor identity
Which operator or person controlled the relevant assets or conduct?
Sponsor or principal
Did an institution fund, direct, task, or control the operator?
Intent
What outcome did the actor seek?
Reach
Who was exposed, and through which channels?
Effect
Did exposure change belief, behavior, policy, turnout, or system performance?
Three separate evidence axes
Status is not confidence.
Confidence is not reliability.
A reliable source can be wrong about one event. A disputed source can provide an authentic document. A likely proposition can still rest on incomplete evidence. Store these judgments separately so one label cannot silently substitute for another.
What the evidence establishes
Claimed · Reported · Corroborated · Substantiated · Proven · Refuted · Unresolved · Superseded
How stable the judgment is
High · Moderate · Low · Undetermined
What is known about the source
Established · Mixed · Unestablished · Deceptive history · Unknown
| Status | Definition |
|---|---|
| Claimed | A named actor or source states the proposition; no independent support is implied. |
| Reported | A publisher or institution reports the proposition with an identified source basis; independent corroboration may still be absent. |
| Corroborated | At least one genuinely independent evidence origin supports the same bounded proposition. |
| Substantiated | Multiple direct and independent lines support the proposition and material counterevidence has been addressed. |
| Proven | A formal adjudication or explicitly equivalent evidentiary threshold establishes the proposition within its stated scope. |
| Refuted | Direct evidence or adjudication materially contradicts the proposition within its stated scope. |
| Unresolved | Available evidence does not support a stable disposition or materially conflicts. |
| Superseded | A later record replaces the current-use status while preserving the earlier claim as history. |
Source-chain independence
Count origins.
Not headlines.
Ten articles that repeat one press statement, actor-controlled post, or unattributed image remain one evidence origin. Independence requires a separate acquisition path, direct record, witness, telemetry source, legal process, or genuinely separate investigation.
Who first created or obtained it?
Record the earliest accessible item, creator role, date, incentives, access, and whether the source is speaking about itself.
How close is it to the proposition?
Separate original records and direct observations from summaries, screenshots, interpretations, and later reporting.
Did another line arise separately?
Copied wording, identical errors, the same crop, or the same anonymous intermediary may expose circular reporting.
Is the item what it claims to be?
Authentication and provenance support origin. They do not automatically prove the truth of the content or the claimed sponsor.
Does it support the same bounded claim?
A second source may corroborate an outage while leaving identity, intent, sponsorship, or effect unresolved.
Can later reviewers reconstruct the path?
Preserve original bytes, hashes, timestamps, transformations, handling events, supersession, and correction history.
Leaderless and pseudonymous networks
A shared label is not
a membership roster.
Symbols, slogans, hashtags, audience exposure, republication, channel administration, campaign organization, technical action, splinter membership, imitation, and external labeling are distinct evidence questions. No role creates movement-wide guilt, loyalty, threat, or authority.
Label or brand
A reusable name, symbol, slogan, or aesthetic that can be adopted without verified authorization.
Cannot infer: Membership, ideology, coordination, or sponsor.
Collective identity
A participant's self-presentation as part of a shared identity or ethos.
Cannot infer: Authority over other participants or permanent affiliation.
Audience
People who view, follow, discuss, or receive the movement's content.
Cannot infer: Support, participation, membership, or susceptibility.
Participant
A person who takes part in a bounded public action or campaign.
Cannot infer: Technical role, command, or responsibility for unrelated conduct.
Amplifier
A person or outlet that redistributes claims or symbols and increases visibility.
Cannot infer: Agreement, coordination, sponsor, or operator identity.
Infrastructure administrator
A person with access to a channel, account, site, document, or other coordination infrastructure.
Cannot infer: Movement-wide leadership or responsibility for all users.
Campaign organizer
A person who proposes goals, wording, logistics, or public calls for a bounded campaign.
Cannot infer: Permanent hierarchy or authority outside the campaign.
Technical actor
A person linked by direct evidence to a specific technical action.
Cannot infer: Ideological representativeness, sponsor, or responsibility for a broader brand.
Splinter group
A bounded group that uses, rejects, or modifies a wider label while maintaining distinct membership or coordination.
Cannot infer: Control of the umbrella label or continuity with every earlier campaign.
Imitator or false flag
An actor that adopts a label or symbols without demonstrated continuity and may seek attention or misattribution.
Cannot infer: Intent or sponsor merely because the claim is false or copied.
Externally assigned member
A person labeled by media, researchers, law enforcement, or opponents rather than by verified self-identification.
Cannot infer: Actual affiliation, guilt, or acceptance of the label.
This taxonomy is descriptive and non-scoring. It must not be used to identify private people, infer protected characteristics, map targets, recruit participants, or establish guilt by association.
Oversight and remedy capability
Compare powers.
Not institutional titles.
A commissioner, court, inspector, tribunal, parliamentary committee, regulator, or ombudsman may each be independent while performing different functions. Describe who can act, what is binding, how a person complains, what secrecy limits disclosure, and how implementation is reviewed.
Ex ante authorization
Can the body approve, reject, narrow, or condition activity before it begins?
Boundary: Authorization is not ongoing inspection or individual remedy.
Records and system access
Can the body obtain relevant files, systems, personnel, and classified material without discretionary withholding?
Boundary: Formal access must be separated from delay, redaction, privilege disputes, and practical usability.
Compelled cooperation
Can the body require answers, production, testimony, preservation, or corrective action?
Boundary: Request, recommendation, and legally enforceable compulsion are different.
Inspection and audit
Can the body inspect live or retrospective practice, sample cases, and test compliance?
Boundary: Periodic thematic review is not the same as continuous operational inspection.
Complaint intake
Can an affected person submit a complaint through a legible and accessible route?
Boundary: Standing, secrecy, jurisdiction, language, disability access, and notice may limit practical use.
Independent adjudication
Can an independent court or tribunal determine legality and facts?
Boundary: Administrative review, parliamentary scrutiny, and adjudication are not interchangeable.
Binding remedy
Can the body stop activity, quash authority, delete data, correct records, compensate harm, or require another enforceable remedy?
Boundary: Criticism and recommendation do not equal binding remedy.
Public reporting
Can the body publish findings, methods, aggregate data, dissent, and implementation status?
Boundary: Publication may be delayed, redacted, filtered, or limited by neither-confirm-nor-deny rules.
Appeal and correction
Can decisions be appealed, errors corrected, and implementation followed over time?
Boundary: A complaint outcome without appeal, implementation evidence, or correction history may not provide effective remedy.
Do not add these capabilities into a country, legitimacy, effectiveness, or rights score. Formal authority, practical access, timeliness, enforceability, accessibility, notification, publication, appeal, and remedy remain separate fields.
Metadata and identity inference
Observed data is not
an identity verdict.
Metadata can support a bounded hypothesis, but joining records, inferring traits, and using an inference in a consequential decision create separate privacy, bias, correction, notice, challenge, and remedy obligations.
Record what the system directly captured: timestamp, identifier, location, transaction, device, document, or event.
Record the transformation, data sources, assumptions, missing data, model version, and uncertainty behind any derived attribute.
Label identity linkage or trait inference as a hypothesis unless a lawful, direct, proportionate verification process establishes more.
Expose who used the inference, for what purpose, with what authority, what consequences followed, and how the affected person can challenge it.
The public method prohibits private-person identification, sensitive-attribute profiling, individualized surveillance, target selection, doxxing, credential discovery, and operational evasion guidance.
Visible correction and supersession
Change the status.
Preserve the record.
A correction is part of verification. Do not silently rewrite an older assertion into new certainty. Keep the earlier wording and evidence state addressable, add the new evidence, explain the change, record publication effect, and expose appeal or further correction routes.
- 01
Preserve the original
Keep the exact claim, claimant, date, evidence, status, and publication context.
- 02
Add the new evidence
Record directness, origin, independence, authenticity, date, scope, and material counterevidence.
- 03
State the disposition
Corroborate, qualify, contradict, refute, withdraw, or supersede only the bounded proposition affected.
- 04
Expose the consequence
Show what wording, page, model, decision, or fictional translation changed—and what remained unresolved.
Human-review readiness
Questions and empty slots.
Not approval.
The 2.0.30 preparation layer gives every source the same eight review categories and every disposition its own empty adjudication slot. Original-language, locally authored, affected-community, privacy, rights, remedy, legal, technical, and editorial fields remain unanswered until genuinely supplied.
Source-specific review questions
Each packet preserves the exact source, wrapper, six prior dispositions, language requirements, local-source needs, and publication blockers.
Append-only adjudication readiness
Claim status, confidence, reliability, attribution rung, origin chain, translation, correction history, reviewer finding, and adjudication remain separate.
Fail closed
No human reviewer, qualified translation, local acceptance, technical attribution, legal approval, completed adjudication, or publication approval is inferred.
Open the human-review readiness method and packet directory.
Individually addressable research custody
Twelve exact sources.
Twelve bounded wrappers.
The preserved reports are research inputs, not a live incident feed or factual certification. Open each source for provenance and each reviewed wrapper for current repository guidance, limitations, safety boundaries, and six explicit publication dispositions.
REVIEWED WITH LIMITATIONS
Information Operations, Disinformation, and Neutral Evidence Presentation
Information-operation vocabulary, behavior-versus-content distinctions, attribution rungs, effects evidence, neutral verbs, and correction practice.
→
REVIEWED WITH LIMITATIONS
Comparative Intelligence Oversight, Transparency, and Remedy
Comparative authorization, review, inspection, parliamentary scrutiny, complaint, adjudication, transparency, and remedy structures.
→
REVIEWED WITH LIMITATIONS
OSINT Verification and Attribution Under Uncertainty
Open-source verification, source independence, authenticity, provenance, confidence language, claim ledgers, and visible correction history.
→
REVIEWED WITH LIMITATIONS
Anonymous in Geopolitical Conflicts and Crisis Information
Anonymous-branded claims during conflicts, verification limits, attribution uncertainty, civilian implications, and differences between symbolic and strategic effect.
→
REVIEWED WITH LIMITATIONS
Anonymous Symbols, Media Representation, and Internal Disagreement
Symbol provenance, media framing, internal fractures, brand appropriation, source laundering, and contested representation.
→
REVIEWED WITH LIMITATIONS
Anonymous Campaigns: Chronology, Claims, Attribution, and Outcomes
Campaign chronology, court-supported attribution, contested claims, correction lists, legal outcomes, and divergence between claimed and documented effects.
→
REVIEWED WITH LIMITATIONS
Anonymous: Origins, Naming, and Early Development (2003–2010)
Origins of the shared name, platform architecture, transition from subculture to activism, contested continuity, and legal framing of digital protest.
→
REVIEWED WITH LIMITATIONS
Anonymous in Courts and Law-Enforcement Records
Court outcomes, evidentiary standards, digital chain of custody, legal definitions, media shorthand, and comparative jurisdictional divergence.
→
REVIEWED WITH LIMITATIONS
Anonymous Sources, Whistleblower Protections, and OSINT Verification Under Uncertainty
Anonymous versus confidential sources, whistleblower protections, corroboration, document authentication, source motive, rights, and editorial failure modes.
→
REVIEWED WITH LIMITATIONS
Anonymous and Other Leaderless or Pseudonymous Movements
Comparative identity boundaries, informal hierarchy, connective action, platform dependence, continuity, infiltration, accountability, and legal labeling.
→
REVIEWED WITH LIMITATIONS
Anonymous: Structure, Membership Claims, and Decision-Making
Role taxonomy, self-identification, informal gatekeeping, campaign-specific authority, splinters, imitators, externally assigned membership, and periodized evolution.
→
REVIEWED WITH LIMITATIONS
Surveillance, Metadata, Data Brokers, and Identity Inference
Commercial metadata, data brokers, re-identification, derived attributes, automated decisions, proxy effects, privacy harms, oversight, and remedy.
→
Browse the preserved source collection, the machine source index, or the 72-disposition review register.
Publication boundary
A method is not
an attribution service.
The release validates custody, structure, schemas, links, and bounded synthesis. It does not establish every source claim, legal conclusion, technical attribution, source identity, translation, causal effect, current event, local acceptance, or political neutrality. It provides no operational cyber, surveillance, coercion, influence-execution, evasion, weapons, targeting, recruitment, or private-person identification guidance.
