In: Statistics and Probability
when describing Type I error under composite null hypothesis, the å usually we select the maximum of the probabilities, can anyone explain this concept for me?
here is an example given: Test H0: µ ≤ µ0 v.s. H1: µ > µ0, then the Type I error is described as å= max Pr[reject H0 | µ≤µ0] = Pr[reject H0 | µ=µ0]
how is this the maximum of the probabilities?