Statistical inference often assumes the normality of the variables involved in the model under study. The existing tests are for independent observations and may not be readily extended to handle the case with correlated ones. In this paper, a transformation method is recommended for normality...