Glossary · Methodology
Statistical significance (p-value)
In short
A result is called statistically significant if it would be unlikely under the assumption of “no effect”, usually at a p-value below 0.05. Significant means neither important nor automatically true.
The p-value answers a narrow question: how surprising would these data be if there were no difference at all? It does not say how likely an effect is, nor how large it is.
In very large studies, even tiny, practically meaningless differences become significant. In small studies, on the other hand, real effects often stay below the threshold. That is why the p-value always needs the effect size and the confidence interval alongside it.
Test enough endpoints and one will eventually come out significant by chance. That is why the pre-specified primary endpoint counts most.
Significant means: probably not chance. Not: large or important.
Where this comes up in the fact checks
- Aspirin for prevention – For the main result p = 0.79, so no sign of a difference; for major bleeding p < 0.001.
- HMB – Muscle strength rose significantly, muscle mass only just at the threshold of significance.
Sources
Related terms: Confidence interval · Effect size · Primary endpoint
Information only, not medical advice. The BK-Score is a subjective assessment according to disclosed rules – not a scientific evaluation. Scoring rules