2. Consider a study with r = 5 treatments and sample sizes ni = 5.
a. Find the values of T, S and B multiples if g = 2, 5 and 10 multiple comparisons are to be made with a 95% confidence coefficient. Which of the three methods performs best here, and which one is the worst? How does the value of g seem to influence the performance of the Bonferroni method?
b. How would the values of T, S and B multiples change if we change all the sample sizes to ni = 20 ? Now do the same comparison as in part a.How do you think the sample sizes influence the comparative performance of the three methods?