Proving once again that people don't know what they think. I remember reading somewhere that researchers were able to phrase questions on a survey such that participants gave opposite answers to the same effective question (too bad I can't seem to find that article on google).