How to read a study
Five questions help separate an interesting finding from a conclusion the study cannot support.
Who or what was studied?
Identify the model, participant group, sample size and follow-up period. A cell experiment, animal study and human trial answer different questions. Results from a narrow population or short follow-up leave other populations and longer-term outcomes uncertain.
What was the comparison?
NIH explains that random assignment to intervention and control groups helps researchers investigate cause and effect. Blinding reduces bias related to knowing a group assignment. A study without a comparable control group is harder to interpret as evidence that an intervention caused a change.
What did the researchers actually measure?
Read the stated outcomes and the reported harms, not only the conclusion. Ask whether the headline describes those outcomes accurately and how the result fits with other studies. Our discussion checklist is: model, comparison, outcome, duration and unresolved questions.
Follow the sources
- 1NIH: Understanding Clinical Studies NIH educational guidance
- 2NIA: Clinical Research, Benefits, Risks, and Safety NIH educational guidance
- 3MedlinePlus: Understanding Medical Research NIH/NLM educational guidance