We’ve already known that in hypothesis testing, we have a null hypothesis, where $ H_0: \theta = \theta_0$ versus alternative hypothesis $H_A: \theta \neq \theta_0$. We also have a presumed pdf function for the variables. Based on these information, we can construct a critical point/region when whatever level of significance $\alpha$ is given. Then if a sample of size $n$ comes in, we can use the sample mean to decide if we accept the null hypothesis or reject it based on whether it falls within the critical region or not.

Read more