Understand a study
Look beyond the headline.
A headline rarely tells the full story. Follow the questions from claim to method, participants, results and practical meaning.
The first version offers selected research explanations and this reading guide. Automatic assessment of arbitrary links or user-submitted claims is not included.
What is the question?
Check what the researchers were actually trying to determine. A study of muscle activation answers a different question from a study of pain reduction or quality of life.
How did they investigate it?
Did they compare treatment groups, follow people over time, or study cells or animals? The method determines what conclusions are justified.
Who took part?
Look for age, life stage, health, country and recruitment process. Who was excluded, and were results reported separately for relevant groups?
How large was the difference?
Look for absolute numbers, comparison groups and follow-up periods. A relative change can seem large when the absolute difference is small.
How uncertain is the result?
Sample size, dropouts, measurement methods and potential bias all matter. Confidence intervals describe statistical uncertainty under the model's assumptions but do not capture every potential error.
How does it fit with other evidence?
Compare the finding with previous studies and systematic reviews. Check that treatments, participants and outcomes are comparable.
A fictional numerical example
“The risk has doubled.” What is the actual difference?
Group A
1 in 100
Experience the event during the hypothetical follow-up period.
Group B
2 in 100
Experience the event during the same period.
The relative risk is twice as high. The absolute difference is one percentage point, or one additional person per 100. Both statements are correct. The figures are fictional and do not describe any actual disease or treatment.
See how we read the evidence.
Can breathing replace pelvic floor exercises?
Distinguishing muscle activation from symptom improvement.
Read the exampleRete ovarii: familiar structure, new evidence
An existing Norwegian-language deep dive distinguishes anatomical knowledge from new mouse research.
Open the original article (Norwegian)Three distinctions we make
Statistical difference versus noticeable improvement
A difference may be statistically clear but too small to affect everyday life. We also examine effect size and clinically meaningful thresholds.
Association versus causation
Two things may occur together because of direct effects, other factors, or data collection. Observational associations alone do not prove causes.
Research finding versus professional recommendation
Recommendations often consider harms, benefits, values and feasibility. We identify the country and organisation issuing them.