9.26 A binomial headache A null hypothesis states that the population proportion p of headache sufferers who have better pain relief with aspirin than with another pain reliever equals 0.50. For a crossover study with 10 subjects, all 10 have better relief with aspirin. If the null hypothesis were true, by the binomial distribution the probability of this sample result (which is the most extreme) equals 10.50210 = 0.001. In fact, this is the small-sample P-value for testing H0: p = 0.50 against Ha: p 7 0.50. Does this P-value give

(a) strong evidence in favor of H0 or

(b) strong evidence against H0?

Explain why.