What is the overall response rate for the 8000 names?
He also notices the sensible fact that in general, the more dollars a person has spent on Luna products, the higher the response rate to the mailing. Yet, the 1000+ category appears to contradict this, by being lower than the value for the 600-999 category.
(Q2): Define p = the true proportion of responses for the 1000+ category on the entire database. Use hypothesis testing, with ? = .05, to test
H0: p ≥ .059 vs. H1: p < .059
In essence, if we accept H0, we can "relax," and not worry that a competitor is targeting our best customers, and that the lower value in the 8000 names (4.9% or .049) is just sample error, and not indicative of a real drop in response rate.