"likely to say they would buy a stun gun" is the Dependent Variable (DV) since that is the outcome measure, the effect caused by the different ads. "More" should not be included since that is a prediction so it is the hypothesis. You don't want to mix your hypotheses and dependent variables because it will cloud your thinking. Also it would be wrong to say "buy a stun gun" is the DV since that was not measured; they only asked for their purchase intentions. Precision is an important goal in learning experimental design since sloppy talking leads to sloppy thinking.