Answer Happy

Q. 1 A scientist wishes to compare K = 3 treatments. They randomize r = 6 subjects to each of these treatments and record the following sample means and variances of the response in each treatment group: УА = 8.2 sa = 3.2 YR = 9.1 = 5.4 Yc = 12.5 sa = 6.1 = si = a) Conduct a test to see if there is evidence of a treatment effect reporting the test statistic, p-value and your conclusions. [8] b) Use the Scheffé method of multiple comparisons to produce 95% confidence intervals for all pairwise differences. [6] c) The response variable is such that a difference of 2 or more units between population means is practically important. What total sample size would be needed if the scientist wanted to have at least 80% power to detect a practically important effect assuming the true variance of the response for any treatment group is 6? [6]