A balanced experimental design has a sample size of n = 11 observations at each of k

Question:

A balanced experimental design has a sample size of n = 11 observations at each of k = 3 factor levels. The sample averages are 1 = 48.05, 2 = 44.74, and 3 = 49.11, and MSE = 4.96.
(a) Calculate pairwise confidence intervals for the factor level means with an overall confidence level of 95%.
(b) What additional sampling would you recommend to reduce the lengths of the pairwise confidence intervals to no more than 2.0?