In practice, we usually do not know the parameters (characteristics) of the whole population. There is only a sample chosen from the population. Point estimators are the characteristics obtained from a random sample. The exactness of the estimator is defined by its standard error. The real parameters of population are in the area of the indicated point estimator. For example, the population parameter arithmetic mean is in the area of the estimator from the sample which is .

If you know the estimators of the sample and their theoretical distributions, you can estimate values of the population parameters with the confidence level () defined in advance. This process is called interval estimation, the interval: confidence interval, and is called a significance level.

The most popular significance level comes to 0.05, 0.01 or 0.001.