MerQuR+Statistics Glossary

Statistics Glossary → Confidence level and confidence interval

Confidence level and confidence intervalconcept

An interval around an estimate that covers the true value with a given long-run success rate; the confidence level is that rate (90%, 95%, 99%).

A mean or difference computed from a sample is a single number and deviates somewhat from the true population value. A confidence interval expresses this uncertainty as a range. A 95% confidence level means: if the same study were repeated many times, about 95% of the intervals computed would contain the true value.

The higher the confidence level, the wider the interval: to be more certain, a wider range must be given. Confidence and significance levels are linked: α = 1 − confidence level. If the 95% interval for a difference does not contain zero, the difference is significant at the 5% level.

An everyday example

When a weather presenter says 'tomorrow's temperature will be between 18 and 22 degrees', they give a range instead of a single number. Widening it to 'between 10 and 30 degrees' would almost never be wrong but would be far less useful; this is the trade-off between the confidence level and the width of the interval.

In MerQur

The 'Confidence level' option in most analyses (90%, 95%, 99%, or a custom value between 80.0 and 99.9 with 'Other'). It changes all confidence intervals and the decision threshold (α) together.

Analyses where this concept appears

Independent-Samples t-TestCox Regression (Proportional Hazards Model)

Related concepts

Bootstrap (resampling)