A crossover experiment with 100 subjects compares two drugs for treating migraine headaches. The response scale is success (1) or failure (0). Half the study subjects, randomly selected, used drug A the first time they had a headache and drug B the next time. For them, 6 had outcomes (1, 1) for (A, B), 25 had outcomes (1,0), 10 had outcomes (0, 1), and 9 had outcomes (0,0). For the 50 subjects who took the drugs in the reverse order, 10 were (1, 1) for (A, B), 20 were (1,0), 12 were (0,1), and 8 were (0,0).

a. Ignoring treatment order, compare the success probabilities for the two drugs. Interpret.

b. McNemar€™s test uses only the pairs of outcomes that differ. For this study, Table 10.15 shows such data from both treatment orders. Testing independence for this table tests whether success rates are identical for the treatments (Gart 1969). Explain why. Analyze these data, and interpret.

Table 10.15:

Treatment That Is Better First Treatment Order Second A, then B B, then A 25 12 10 20 20