Suppose a researcher wishes to test H0 U = 100 against H1U>100 using 0.05 level of significance; however if she obtains a sample mean far enough below 100 to suggest that is Ho unreasonable, she will switch her alternative hypothesis to Ho not equal 100 (significance level 0.05 with the same sample data. Assume H- to be true. What is the probability that this decision strategy will result in a type I error. Sketch the sampling distribution and put in the region of rejection?