Tests to determine if the sample data follows a normal distribution.
There are four types of normality tests depending on the number of variables selected.
Normality Test
Description
Anderson-Darling
Anderson-Darling (parameters assumed unknown): Select this option to compute the Anderson-Darling test of normality. The p-value is computed under the assumption that the parameters are unknown.
Anderson-Darling (parameters assumed known): Select this option to compute the Anderson-Darling test of normality. The p-value is computed under the assumption that the parameters are known.
Kolmogorov-Smirnov test
KS test (parameters assumed unknown): Select this option to compute the Kolmogorov-Smirnov test of normality. The p-value is computed under the assumption that the parameters are unknown.
KS test (parameters assumed known): Select this option to compute the Kolmogorov-Smirnov test of normality. The p-value is computed under the assumption that the parameters are known.
Shapiro-Wilk
Select this option to compute the Shapiro-Wilk normality test. This test is available for n up to 5,000.
Mardia multivariate normality test
If two or more variables are selected, then select this option to compute the Mardia multivariate normality tests based on multivariate skewness and kurtosis.
