"Statistically significant" tells you that an observed effect is unlikely to be due to chance. It does not tell you how large the effect is, whether it matters in practice, or whether a treatment works for you. Statistical significance is not the same as clinical importance, and a result can be statistically significant without being important the BMJ analysis of what is wrong with significance tests.
Key takeaways
- Statistical significance concerns chance; clinical importance concerns real-world meaning.
- A statistically significant result can still be too small to matter.
- A non-significant result does not prove that no effect exists.
- Estimates and confidence intervals give more useful information than a single significance verdict.
- This page explains the distinction; it does not evaluate any treatment.
What readers should know
A significance test answers one narrow question: could this difference be due to chance? The BMJ analysis argues that relying on significance tests alone misleads interpretation and that estimates and confidence intervals should be used instead. When you read "statistically significant," ask what the size of the effect was and whether it would matter to the people affected, because the label alone is not a measure of importance.
Evidence boundaries
- Statistical significance is not evidence of efficacy, clinical importance, or causation.
- A "significant" result does not mean a treatment works for everyone.
- A "non-significant" result does not mean nothing happened.
- See microneedling evidence limits for how evidence limits are presented on specific topics.
FAQ
If a result is statistically significant, does that mean the treatment is effective? No. It means the difference is unlikely to be chance; it does not tell you whether the difference matters.
If a result is not statistically significant, does that mean the treatment has no effect? Not necessarily. The study may have been too small to detect an effect.
What should I look at besides significance? The size of the effect and the confidence interval around it.