摘要: Testing for multivariate normality in a sample of observations presents greater difficulty than testing univariate normality. Tests are mostly based on either the unique properties normality, generalizations tests or goodness-of-fit. Several methods assessing presented here, with focus those that easily applied and/or have good power properties. Some plots can be used subjectively determining also described. Keywords: squared radii; Q–Q plots; multivariate skewness; multivariate kurtosis; regression tests; correlation tests; Shapiro–Wilk test