Problem 3. (12 points) We collect the data of the differences between midterm scores and final scores (both based on scale) of each student in two sections. A positive difference means that the student's final score is higher than the midterm score. We want to test whether the students of section A shown more improvement than the students of section B. Assume significant level a = 0.05 Section A 12 8 14 15 11 10 9 10 -70 11 Section B 7 4 7 -21 -18 5 -2 15 (a) (6 points) Compute the statistic for Mann-Whitney U-test. (b) (4 points) Assume normal approximation, compute the approximate p-value for the observed test statistic. (c) (2 points) What is the advantage of using Mann-Whitney U test over a para- metric test for this