When the means of samples drawn from normal populations are employed in performing tests of hypothesis, the experimenter usually assumes that the sample was drawn from a complete normal population.