Question

1. A randomized controlled trial was performed to see if patients with dyslipidemia could improve their lipid profile with a regular exercise regime. Participants were randomized to one of 3 different exercise programs or control. The lipid profiles of participants were assessed at study entry and 8 months later. Data for HDL levels are presented in the table below. Assume that HDL levels have a distribution fairly close to the Normal distribution.

Table 1: HDL levels (Mean ± SD) at study entry and after the exercise programs
Study Group	Study entry	After exercise	Change
Control (n=26)	42.7 ± 2.7	42.1 ± 2.2	-0.6 ± 2.4
Low amount moderate intensity (n=19)	40.3 ± 2.2	41.0 ± 2.6	0.7 ± 2.7
Low amount high intensity (n=17)	46.6 ± 3.7	46.9 ± 3.4	0.3 ± 3.5
High amount high intensity (n=22)	44.3 ± 2.9	48.6 ± 3.3	4.3 ± 3.0

1. *(1 point) Suppose now that the investigators wanted to do a separate test for each group (control, low amount moderate intensity, low amount high intensity, high amount high intensity) for whether there is evidence of on average a change in HDL before and after the intervention. What would be an appropriate significance level to use if one wanted to assure that the family-wise error rate was no larger than 5%?
2. From the data presented, can you calculate a 95% confidence interval for the difference in the change in HDL between the High exercise and the control group? If yes, calculate it. If not, what more information would you need?

Answer 1

B) From the data presented ,we cannot calculated a 95% confidence interval for the change in HDL between study entry and after exercise in the high exercise group. because we need sample size to calculate confidence interval.

NOTE: ACCORDING TO MY KNOWLEDGE I HAVE ANSWERED ONLY (B) .I AM LITTLE BIT CONFUSE AT (A).THANKS.