the answer is D.
In only polling people in Arizona, you aren't taking a representative sample of the total population. You want to represent all *U.S.* adults.
What if people from Arizona spend proportionally more (or less) than the rest of the nation?