MAKING IT USEABLE    Interpreting Estimates       Calculating Confidence Limits on Sample Statistics

Confidence limits are also an important part of the assessment of error and are determined based on how much variability there is in the data and how sure or unsure one is of the value obtained from the sample.

As the size of the dataset increases, so the confidence limits become closer and closer to the values estimated from the data.

Usually, one computes either the 95 percent or 99 percent confidence limits. These limits indicate respectively the range of values within which one has either 95 percent or 99 percent confidence that the true value of a population statistic will lie, based on the estimate from the sample. Therefore, 99 percent confidence limits define a wider range than 95 percent confidence limits. For example:

• An interpretation of 95 percent confidence limits is that one has a smaller than 1 in 20 chance that the true population value will be outside this range.
• Similarly, an interpretation of 99 percent confidence limits is that one has a smaller than 1 in 100 chance that the true population value is outside this range.

---