Abstract: Normal distribution is one of the important distributions for continuous variables. In many statistical analyses, researchers often assumed directly the data were normally distributed without test before hand. If the numerical variables are not normally distributed, it may lead to wrong results. Normality tests which are commonly used can be classified into two categories. One category includes subjective graphical methods, and the other category contains computational methods including hypothesis tests which are objective and quantitative. In this paper, several normality tests popularly used will be introduced according to the above two categories to provide references for proper applications of statistical analyses of numerical variables.