A confidence interval indicates how uncertain a researcher is about an estimated range of values. A 99 percent confidence interval indicates that if the sampling procedure is repeated, there is a 99 percent chance that the true average actually falls between the estimated range of values.
Confidence intervals allow researchers to describe how stable an estimate is. A lower confidence interval, for example an interval of 90 percent, implies a less accurate estimated range of values. Selecting a percentage for the confidence interval is not set in stone and often changes from one discipline to another. Any percentage can be used when setting a confidence interval, but the most common confidence interval percentages are 90 percent, 95 percent or 99 percent.
To calculate a confidence interval, a researcher must have three pieces of information: confidence level, statistic and margin of error. The confidence interval is derived from the sample statistic plus or minus the margin of error.
In statistics, confidence intervals are closely tied to significance levels. For example, a 99 percent confidence interval is equivalent to a significance level of 0.01, or 1 percent. When a statistic is significant at the 0.01 level, researchers conclude that the chance that the statistical significance occurred due to chance is less than 1 percent.