An enthusiastic junior executive has run a test of his new marketing program. He reports that it resulted in a “significant” increase in sales. A footnote on his report explains that he used an alpha level of 7.2% for his test. Presumably, he performed a hypothesis test against the null hypothesis of no change in sales.
a) If instead he had used an alpha level of 5%, is it more or less likely that he would have rejected his null hypothesis? Explain.
b) If he chose the alpha level 7.2% so that he could claim statistical significance, explain why this is not an ethical use of statistics.