A reference beside a supplier statement is a starting point, not the end of a source check. The linked page may be a review, a different experiment or a paper that supports a narrower statement. A useful check follows the claim back to the original observation and compares the two precisely.
Write down the claim before searching
Copy the exact statement into your private research notes with its URL and access date. Then separate its components: which material is named, what effect is asserted, in which system, compared with what and over what period? If the wording leaves these details unspecified, mark them as missing.
For an original hypothetical example, “peptide X improves repair” is much broader than “a specified cell assay showed an increase in a measured migration endpoint under the reported conditions”. Before judging support, you need to establish what “repair” refers to in the source.
This first step prevents the search from quietly changing the question. Finding any positive-sounding paper is easier than finding evidence for the exact statement. Keep the original wording visible while you investigate.
Use the citation's identifiers
If a title, author, year, DOI or PMID is supplied, use those details to locate the record. PubMed supports citation searching and searching by PMID. Its guide also explains field-specific searches when a name or phrase needs narrowing. Verify the returned title and bibliographic details rather than relying on a familiar keyword.PubMed search guidance (opens in a new tab)
A DOI or PMID helps identify a document; it does not establish that the document supports the supplier's sentence. If the supplied reference leads somewhere unexpected, preserve that observation rather than silently replacing it with a more convenient paper.
Distinguish the original experiment from commentary
Open the available article and identify its type. A review may discuss many experiments without being the report that generated the particular result. Follow the relevant reference to the original study where possible. Keep both records if the review helped you locate it, but distinguish their roles in your notes.
Then find the methods and results supporting the statement. An introduction can describe background expectations; a discussion can propose a mechanism; neither is automatically a measurement made in that study. Locate the actual experiment and the figure, table or result passage on which the claim depends.
If only an abstract is accessible, label the check as abstract-only. Do not claim to have checked methods or supplementary details you could not read. An accessible summary may establish the study's broad scope while leaving important qualification unresolved.
Compare the claim and evidence side by side
| Question | What to record |
|---|---|
| Same material? | Sequence, modification, formulation or other identity detail |
| Same system? | Cell model, animal model, human study or analytical experiment |
| Same comparison? | The actual control or comparator |
| Same outcome? | The measurement made, not a broader interpretation |
| Same conditions? | Relevant exposure, timing and experimental context |
In the hypothetical repair example, a migration assay does not automatically establish tissue repair in another system. A narrower finding may still be scientifically interesting. The correction is to describe that finding accurately, not to dismiss the entire paper or stretch its conclusion.
Also check whether the paper contains a correction, retraction notice or later linked clarification. Record the document version and the nature of any notice. A source check should follow the current record, not only a saved screenshot of an earlier headline.
Leave a short evidence trail
- Original statement, location and access date.
- Resolved citation and document type.
- Exact experiment or result supporting the interpretation.
- Differences between the claim and the study's scope.
- Access limitations and any unresolved questions.
A useful conclusion can be “the paper supports a narrower result in this model” or “the cited source does not contain the stated experiment”. The goal is a traceable judgement, not a binary label based solely on whether a reference exists.
Sources and further detail
- NLM — PubMed user guide (opens in a new tab)
Citation searching, Advanced Search, History and Search Details. Database behaviour was checked on the access date.
Sources checked 19 September 2026. Worked examples are illustrative unless a supplied report is explicitly identified. This article has not undergone independent scientific peer review.