Who are they fooling with this one?
Isn’t there a confounding factor here? Couldn’t it just be that the books participating are often recent, well-marketed books from major publishers? Those are bound to be more popular than older, less publicized ones. The more interesting comparison would have been between sales of the 120,000 books participating in the program over time i.e. before the program vs. now. In essence, there are two questions here: Amazon asked, “What difference is there in sales of books in the program and books not in the program?” when they should have asked, “What effect does Search Inside the Book have on sales?”