A Modified Procedure for a One-Sided Normal Test

The purpose of this article is to propose a modified procedure for a one-sided normal test. This modified procedure is different from the conventional one in that it uses sample information for establishing the null and the alternative hypotheses. We show that the power of the modified procedure is greater than that of the conventional one in all practically possible statistical situations. In spite of the greater power of the modified test procedure it neither requires significantly more computation nor is more complicated in procedure than the conventional one. Therefore the modified procedure may be said to be a significant methodological improvement over the conventional test procedure.