Evidence-appraisal glossary
Credible interval
A credible interval is a Bayesian range that, given the data and a prior, contains the unknown parameter with a stated probability (say 95%). Unlike a confidence interval, you can literally say there is a 95% probability the true value lies inside it.
Also called: Bayesian credible interval, Bayesian confidence interval, posterior interval.
What it is
A credible interval comes from Bayesian analysis. After combining the observed data with a prior (what was assumed before the study), you get a posterior distribution for the parameter. The credible interval is the slice of that posterior holding a stated probability, commonly 95%.
It answers the question people wrongly attribute to a confidence interval: "given this evidence, where does the true value probably sit?" A 95% credible interval genuinely means a 95% probability the parameter falls in that range.
How to use it when reading a study
- Check the prior. The interval depends on it. A strong or optimistic prior can narrow or shift the range, so look for how the prior was chosen and whether the authors tested alternatives (a sensitivity analysis).
- Note the type. Equal-tailed intervals cut 2.5% off each end; highest-density intervals (HDI) take the most probable values. They differ for skewed posteriors.
- Read the width as uncertainty, and see whether it crosses a no-effect value.
This is a plain-language methodology definition for reading research. It is general education, not medical advice.