Observational studies
Researchers watch what happens without changing anything — for example following thousands of people for years (a cohort study), or comparing people with and without a disease (a case-control study). They can find links, but a link is not proof of cause: something else may explain both.
Randomised controlled trials
People are assigned by chance to a treatment or a comparison group. Because chance decides, the groups are alike in other ways, so differences in outcome are more likely to be caused by the treatment. Trials are the standard way to test treatments.
Systematic reviews and meta-analyses
These collect every good-quality study on a question and weigh them together. They usually give the most reliable answer.
Other things to check
- Who was studied — people, animals or cells? Results in mice or in a lab dish may not apply to people.
- How many — larger studies are less affected by chance.
- How big the effect — see relative vs absolute risk.
- Peer review — whether other experts checked the work before publication.
Sources
- Understanding Medical Research — MedlinePlus — U.S. National Library of Medicine. Checked 2026-09-26.
Editorial update: 2026-09-26.
Medical disclaimer: this is a source-linked educational summary, not individual medical advice, diagnosis, prescription or interpretation of personal results. For severe or rapidly worsening symptoms, call 115.
Antopeloni — plain-language health education from public sources, part of the ZIBADIS network founded by Masoud Moghaddam.
