|
Abstract:
|
In statistical inference, oftentimes we assume that the data are coming from normal distributions. This is due to the fact that the majority of parametric statistical theories have been developed based on normal theory. Therefore, testing the validity of the normality assumption is a key part in such statistical data analyses. In this study, we will investigate ten of currently available test of normality methods by using a monte-Carlo simulation. Alternative distributions will be used to calculate the empirical power of the tests. These alternative distributions were used from three different categories namely symmetric, asymmetric short tailed and asymmetric long tailed distributions. In addition to this, we will also study the performance of the methods in the presence of outliers.
|
