O dokładności analitycznego wyznaczania mocy pewnego testu na normalność rozkładu prawdopodobieństwa
On estimation of the power of a normality test

Summary/Abstract: The test of normality based on rank correlation between the sample mean and the sample variance is considered. The well known Geary’s theorem leads to the conclusion that such a correlation is not significant when the hypothesis on normality distribution is true. The paper considers the estimator of the power of the test in the case of asymmetric alternatives. On the basis of computer simulation analysis the accuracy of this estimator is analysed. The chi-square and Dagum’s distributions are considered as alternative ones. The analysis of the simulation results leads to a conclusion that the proposed estimator of the power can be treated as sufficiently accurate from the practical point of view.

• Details
• Contents