Best Critical Region: Let be a critical region for a hypothesis test with significance level, where the hypotheses are and . Then we say that is a best critical region of size if whenever is another critical region with size , then .

What this means is that if the alternative hypothesis is true, then the probability that we reject the null hypothesis is greatest if we use the critical region . Besides, this best critical regions is the most powerful among all the regions with size , and thus the UMPT.

Neyman-Pearson Lemma: Let be a random sample sample from a distribution with PDF , and let the likelihood function . If there exists a positive constant and a region such that

  1. for and for Then is a best critical region of size for testing against

Example: Let be a random sample from .

  1. Show is a best critical region for testing against .
  2. Find and so that and .

First we calculate

We need this ratio to be less than a positive constant

for , so we can just let .

We want and , that is

where is the standard error, and . Then we could solve for .