A method for testing whether a specified probability distribution resulted in a given data sample from a population. This test is considered an effective way of determining whether the distribution of data in a sample departs from a normal distribution. The Anderson-Darling normality test was developed by Theodore Anderson and Donald Darling in 1952.