From Wikiversity
Confidence intervals (CIs) indicate the reliability of a statistical parameter. Rather than rely on a single estimate for a parameter (such as a measure of central tendency, a CI approach involves specifying a likely range in which the central tendency for a variable of interest for a particular population would lay.
For example, the mean height for a sample of adult human beings might be, say, 170cm. However, from the sample we may conclude that the mean height for the entire population of adult human beings is 95% likely to be between 168cm and 172cm (a lower and upper confidence level respectively).
[edit] See also
