A certain brand of phone is supposed to weight 25 grams. We assume that the phone weight follow a normal distribution with the standard deviation off s=3 grams.We buy 4 such phones and compute the mean weight in order to test H0 : u 25 g vs H1: u<25 g. If we define our critical region to be c: xbar<22.5, what is the probability of a type I error of our test?