Procedure that relies on finding a rule to decide for which value range of statistic based on random sample the zero-hypothesis should be rejected and alternative hypothesis accepted on the given significance level.