What a Useful Research Reference Needs

The main name and aliases should point to the same entity. Fragments, variants, fusions and mixtures need explicit distinctions. NIH authentication of biological and chemical resources provides a basis for this priority: materials central to research require appropriate identification and verification.

As an editorial application, the reference should separate confirmed information from unavailable information. Adding a formula or sequence because of a similar name can make the text more detailed and, at the same time, less reliable.

ARRIVE organizes information needed to interpret animal research, including groups, the experimental unit and animal characteristics. These elements help establish what an observation actually represents.

In a reference, the model should appear close to the result, not only in a distant note. The comparator should also be identified: vehicle, unexposed control or another entity. Without that information, a difference loses the reference point that gives it meaning.

The same guideline distinguishes outcomes, analysis and results with their variability. A brief synthesis can preserve that structure without reproducing the whole article: what was measured, how it was compared and what uncertainty accompanied the observation.

This selection avoids replacing a specific readout with an overly broad category. A molecular marker, a histological image and a behavioral task can belong to the same study, but they are not equivalent results. The reference should name the one supporting its claim.

NLM links original records to corrections, retractions and other notices. A bibliographic check does not end when the PMID matches the title: those relationships must also be reviewed.

A correction can concern funding, text or figures, and its scope should be described precisely. Finding a notice does not justify declaring the entire study invalid; ignoring it does not justify treating a corrected version as current. The reference's review date helps place that check in time.

A reference can state which information is missing without turning that absence into repetitive content. The useful approach is to identify a specific consequence: the variant was not confirmed, the mixture was not examined or the source supports only a description of the model used.

This is a proposal for editorial organization, not a certification granted by NIH, ARRIVE or NLM. Its purpose is to connect material, experiment and document so readers can examine a claim and identify where its support ends.