Sources and Citation Policy
The Data Verdict uses sourcing and citation to help readers understand, evaluate, and independently examine the evidence behind our journalism.
Purpose
Sources provide the evidentiary foundation for journalism.
The Data Verdict seeks to identify and appropriately attribute material evidence wherever doing so is compatible with accuracy, safety, confidentiality, applicable law, and the public interest.
Citation is not decoration. A source should genuinely support the claim for which it is cited.
The level of sourcing and verification required increases with the seriousness, novelty, uncertainty, potential harm, and public impact of a claim.
Claims involving serious allegations, public safety, legal matters, scientific findings, financial information, security incidents, or a person’s conduct require particularly careful sourcing and verification.
Source Hierarchy and Assessment
Where available, we generally prefer direct, primary, and contemporaneous evidence.
This may include:
- Laws and regulations
- Court records and judicial decisions
- Official statistics
- Regulatory decisions
- Audited financial filings
- Contracts and formal agreements
- Technical documentation
- Peer-reviewed research
- Research datasets
- Interviews with relevant sources
- Authenticated correspondence
- Original photographs and video
- System or technical records
- Direct observation
Primary evidence is not automatically reliable.
A company filing may omit relevant context. A government statement may reflect institutional or political interests. A technical demonstration may be controlled or staged. A witness may be mistaken. A dataset may contain methodological limitations or biases.
Editors therefore assess factors such as:
- Provenance
- Authenticity
- Independence
- Expertise
- Incentives and potential conflicts
- Methodology
- Date and relevance
- Completeness
- Corroboration
Secondary sources can provide valuable context, interpretation, comparison, and reporting leads.
However, a claim does not become independently verified simply because it has been repeated by multiple publications. We seek to avoid circular sourcing, untraceable statistics, and chains of reporting that ultimately rely on the same unsupported source.
Attribution
Material obtained from other sources should be attributed in a manner appropriate to the format and circumstances.
This may include attribution of:
- Facts
- Quotations
- Data
- Distinctive analysis
- Images
- Documents
- Research findings
- Statements from interested parties
Where practical, online citations should link directly to the supporting document, record, dataset, study, statement, or other relevant source rather than to an unrelated homepage or generic search result.
Quotations must preserve the speaker’s meaning and material context.
Minor edits for clarity, such as removing verbal fillers or obvious transcription artifacts, may be made when they do not alter the meaning of the statement.
Translations should be identified where relevant. Particular care should be taken with disputed, technical, culturally sensitive, or legally significant language.
Links, Archives, and Access Dates
When a source is publicly available, we seek to link to an authoritative and reasonably stable version where practical.
For information that may change, disappear, or become inaccessible, we may preserve a lawful archive, screenshot, downloaded copy, or internal verification record for editorial purposes.
Preserving a copy for verification does not grant permission to republish copyrighted material beyond what is permitted by applicable law.
Publication dates, update dates, access dates, retrieval dates, or archival capture dates may be provided when they are material to understanding the source.
If a cited source becomes unavailable, a broken link may be replaced with an official archive, stable version, or substantially equivalent authoritative source where appropriate.
Material changes to sourcing should be recorded through the relevant editorial process.
Data and Research
Data-driven claims should identify, where relevant:
- The source of the data
- The relevant time period
- The population or sample
- Units of measurement
- Methodology
- Significant assumptions
- Material limitations
We do not describe correlation as causation without sufficient evidence to support a causal conclusion.
Particular caution is applied to:
- Small or non-representative samples
- Self-selected surveys
- Unpublished methodologies
- Vendor-funded research
- Preprints
- Model-generated estimates
- Data with significant missing values or known limitations
Research findings should be described in proportion to the strength of the evidence.
A single study is not presented as scientific consensus merely because it supports a particular conclusion.
Where material to interpretation, we consider factors including retractions, corrections, conflicts of interest, peer-review status, methodological limitations, and replication concerns.
Company, Government, and Advocacy Materials
Press releases, investor presentations, product demonstrations, corporate white papers, commissioned research, government statements, advocacy reports, campaign materials, and similar publications can contain useful information.
However, such materials are generally interested sources and should be identified as originating from the organization or party with an interest in the subject.
Important claims should be independently tested where reasonably feasible.
Where a company, founder, government, political campaign, advocacy organization, or other interested party provides substantial documents, data, interviews, access, travel, research, or editorial material, the relationship should be disclosed when it would reasonably help readers evaluate the resulting coverage.
A claim does not become an established fact merely because an interested organization has supplied it or because The Data Verdict has reported that the organization made the claim.
Anonymous and Confidential Sources
Anonymous sourcing may be used when:
- The information has meaningful public-interest value.
- The source has a credible reason to require confidentiality.
- The information cannot reasonably be obtained on the record.
- The responsible editor knows, or has sufficient grounds to verify, the source’s identity.
- The benefits of publication outweigh the risks associated with anonymity.
Where safety permits, an article should explain why a source has been granted anonymity.
A single anonymous source will rarely be sufficient to support a serious or contested allegation.
Corroboration may include documents, independent witnesses, data, direct observation, technical evidence, or other credible sources.
Anonymity should not be granted merely to facilitate an avoidable personal attack, promote an undisclosed private interest, or allow someone to avoid accountability for a routine opinion or claim that could reasonably be made on the record.
Confidential sources should be handled according to the publication’s source-protection and editorial standards.
User-Generated and Social-Media Material
Posts, screenshots, direct messages, photographs, videos, audio recordings, comments, and other user-generated material may be useful evidence, but online availability or virality does not establish authenticity.
Before publication, relevant material should be verified proportionately to its importance and risk.
Verification may include:
- Reviewing original files
- Examining available metadata
- Checking chronology
- Geolocating visual material where possible
- Reviewing account history
- Conducting reverse-image or reverse-media searches
- Contacting the original source
- Comparing material with independent evidence
- Examining visual or audio characteristics
- Checking whether material has been edited or taken out of context
We seek to minimize unnecessary republication of personal information, graphic content, private communications, or other material that could create avoidable harm.
The fact that information is publicly accessible does not eliminate privacy, safety, consent, or contextual considerations.
Artificial Intelligence and Synthetic Material
Artificial intelligence systems may assist with research discovery, transcription, translation, data processing, organization, or other editorial workflow tasks.
However, AI output is not itself treated as a source or evidence.
A model-generated answer, summary, citation, quotation, image, transcription, or factual claim must be checked against appropriate original material before being relied upon in published journalism.
We do not knowingly publish AI-invented references or treat a model’s confidence, reasoning, or apparent certainty as evidence of accuracy.
Synthetic or materially altered media should be identified when its artificial or altered nature is relevant to the reader’s understanding.
Claims concerning deepfakes, cloned voices, generated images, manipulated documents, or other synthetic media require technical and contextual verification proportionate to the potential consequences of getting the determination wrong.
Source Records and Protection
We retain source records for as long as reasonably necessary to support verification, corrections, legal obligations, security, and editorial accountability, subject to privacy, confidentiality, and source-protection considerations.
Access to sensitive source material should be limited according to legitimate editorial need and the sensitivity of the information.
Promises of confidentiality should be made carefully and only by an appropriately authorized editor.
No journalist or publication can guarantee absolute protection against every possible legal demand, device compromise, security incident, technical failure, or other external event.
Where material limitations on confidentiality or security are known, sources should be informed where reasonably possible.
Corrections and Challenges
A citation or source reference should be corrected when it:
- Does not actually support the associated claim
- Identifies the wrong source
- Links to the wrong document or version
- Omits material context
- Mischaracterizes the source
- Becomes materially misleading because of a subsequent change
Readers, researchers, cited organizations, and other relevant parties may raise sourcing concerns through:
corrections@thedataverdict.com
A challenge to a source does not automatically establish that the published claim is incorrect. The editorial team will assess the underlying evidence and make an appropriate correction, clarification, update, or other response where warranted.
Last updated: 25 August 2026.