Examples are the skewness test, the kurtosis test, the D’Agostino-Pearson omnibus test, the Jarque-Bera test.
Tests based on descriptive statistics of the sample. Examples are the Kolmogorov-Smirnov test, the Lilliefors test, the Anderson-Darling test, the Cramér-von Mises criterion, as well as the Shapiro-Wilk and Shapiro-Francia tests. Tests based on comparison (“best fit”) with a given distribution, often specified in terms of its cumulative distribution funtion (cdf). Tests to evaluate normality or a specific distribution, frequentist approaches can be broadly divided into two categories: One area where (nowadays) simple computer runs can be illuminating is with respect to normality tests. 
We are now at a much more comfortable position to analyse how these methods perform. Even in the 1970’s and 1980’s, when computers were already available at Universities and even at home, certain matrix operations were limited by hardware constraints. Before the fast computers era, it was difficult to test how valid certain statistical methods were.
0 kommentar(er)
