Editorial policy

Evidence before opinion

Evidence has priority over confident language. A statement should not become more certain than the material that supports it.

• Factual claims should be traceable to documentation, direct observation, or a clearly identified external source.
• We identify manufacturer claims as manufacturer claims rather than presenting them as independent verification.
• We attribute independent findings to the publication, laboratory, reviewer, clinician, standards body, or other source that produced them.
• We keep material uncertainty, limitations, and conflicts in the evidence visible.
• We label editorial inference and do not present it as independently verified fact.
• We do not describe research as hands-on or laboratory testing when that work did not occur.

These rules apply to both the conclusion and the supporting detail. A useful recommendation should make its evidentiary basis understandable without implying certainty that the available sources cannot support.

Editorial independence

Editorial decisions are made separately from affiliate commission outcomes. Affiliate revenue does not determine product inclusion, evaluation criteria, criticism, ranking order, or recommendation.

ChoiceCheckLab does not sell placement or operate a pay-to-rank model. Manufacturers, retailers, and affiliate programs do not receive control over editorial conclusions.

Commercial availability may affect whether a product remains practically useful to recommend. A discontinued product, an unavailable model, or a listing that no longer identifies the intended variant can change the usefulness of a recommendation.

That is an availability judgment, not a commission judgment. Commission rate does not determine editorial ranking or recommendation.

Source standards

Sources are selected for the claim they can support. A product listing can identify a current retail variant, while a manual, technical document, independent test, or specialist source may be better suited to a different question.

• Primary sources — We use official manufacturer documentation to verify exact product identity, specifications, supported features, limitations, and warranty terms where appropriate.
• Independent sources — We use reputable independent testing, specialist reviews, and clinical or technical sources when they add relevant evidence beyond manufacturer material.
• Attribution — Third-party findings remain attributed. ChoiceCheckLab does not present external testing, measurements, or conclusions as its own work.
• Product identity — Specifications, images, retailer destinations, and conclusions should refer to the same exact model or to a clearly identified variant.
• Conflicting evidence — Material conflicts between reliable sources remain visible rather than being silently resolved in favor of the more convenient claim.

The existence of a source does not make every claim equally strong. We consider what the source actually measured, documented, observed, or asserted, and keep those distinctions visible in the article.

AI-assisted work

AI tools may assist with structure, controlled drafting, information organization, consistency checks, technical or code work, and quality assurance.

AI output is not, by itself, evidence, source verification, expert testimony, or product testing. It is not justification for invented specifications, compatibility claims, measurements, credentials, or outcomes.

Claims still need to be checked against the material that supports them. Unsupported wording should be removed or qualified regardless of whether it was drafted by a person or an AI tool.

Editorial responsibility remains with the named editor. AI-assisted material is reviewed before publication.

Corrections and updates

A correction addresses a factual or attribution error in existing content. Examples include an incorrect specification, the wrong ASIN or model identity, a compatibility error, incorrect attribution, or an unsupported factual statement.

An update responds to a material change in the product or evidence environment. Examples include a replacement model, a discontinued product, a material availability change, a safety recall, new credible independent evidence, or a meaningful change in manufacturer documentation.

Material corrections or updates may trigger editorial review of the surrounding comparison and its conclusion. We do not claim a fixed review schedule, and a minor wording edit does not automatically mean that the full buying decision has been re-researched.

When a reader identifies a possible error, the most useful report includes the page address, the statement in question, and a relevant source where one is available.