Identify the evidence type
Ask whether the source is a laboratory experiment, animal study, observational report, randomised clinical trial or review. Each supports a different level of conclusion.
Read the methods and measured endpoint, not only the headline. A statistically detectable difference may still be small, indirect or uncertain.
Look for confirmation
Check sample size, comparison group, follow-up period, publication status, funding and conflicts. One study is a signal to investigate, not a final answer.
PubMed helps locate indexed literature; ClinicalTrials.gov records registered studies and their status. Registration itself does not prove effectiveness.
Keep a claim-to-source record
For a percentage or timeline, save the original publication, material/form, model, sample size, comparator, endpoint, follow-up and exact result. If one of these is missing, record that gap instead of adding detail from a different experiment.
A search-results page is a discovery tool, not the supporting citation. Link the exact paper or registry record and distinguish its publication date from the date you checked it.
Compare like with like
Check whether a study uses the same sequence, form or constituents as the product record. A full-length peptide, a fragment and a blend require separate evidence notes; shared names do not establish equivalence.
Separate selected example papers from a systematic review. Record the search terms, date and inclusion rules if you publish a measured literature count or ranking.