A research laboratory was developing a new compound for allergy relief, and was manipulating the dosage of two ingredients. Information on the dataset follows: Column 1: Relief: The number of hours of relief.
Column 2: DoseA: The dosage of the first ingredient, with values of Low, Med, High. Consider this to be factor A.
Column 3: DoseB: The dosage of the second ingredient, with values of Low, Med, High. Consider this to be factor B.

You may assume that i 1 is low, i 2 is medium, and i 3 is high, and similarly for j. The following table gives the group means, and the sample sizes (in parentheses): Further, consider the appropriate model to use the model with

Low Med High A
Low 2.48(4) 4.60(4) 4.58(4) 3.88(12)
Med 5.45(4) 8.93(4) 9.12(4) 7.83(12)
High 5.97(4) 10.28(4) 13.25(4) 9.83(12)
B 4.63(12) 7.93(12) 8.98(12) 7.18(36)

interactions, where SSE = 1.63.

Required:
a. Describe the overall trend for factor A (regardless of interactions) in words.
b. Describe the overall trend for factor B (regardless of interactions) in words.
c. If you had to build 6 pairwise confidence intervals, which correction would you use (Tukey, Scheffe, Bonferroni)? Explain your answer.
d. Which factor level (of either A or B) would you suggest is the most important for increasing the number of hours of relief? Justify your answer.