Browse the glossary:
CI, margin of error,
A confidence interval reflects the extent to which the play of chance may be responsible for a result from a study, such as an effect estimate (or the accuracy of a diagnostic test or the opinions of a population).
A 95% CI means that we can be 95% confident that the actual size of the effect is between the lower and upper number given (confidence limit). This means there is a 5% chance that the actual effect is outside of this range.
Wider confidence intervals indicate lower precision of the effect estimate; narrower intervals, greater precision.
An effect estimate (or other study result) can be precise (i.e. have a narrow confidence interval) but biased (i.e. distorted away from the actual effect because of systematic errors).
Occasionally other intervals are reported, such as 90% or 99% confidence intervals.
If you feel that this definition hasn't helped you to understand the term, click on our monkey to let us know.