Dan Froomkin has a good example of the importance of knowing how poll questions are worded before deciding on what the percentages for and against something mean:
Call it a tale of two questions.
A Gallup/USA Today poll finds a clear majority -- 57 percent -- of Americans supporting the withdrawal of U.S. troops from Iraq; while a Washington Post/ABC News poll finds only a narrow minority -- 47 -- percent in favor.
How can that be?
Well, look at the wording.
Here's the Gallup question: "Which comes closer to your view? Congress should pass a resolution that outlines a plan for withdrawal of U.S. troops from Iraq (or) decisions about withdrawing U.S. troops from Iraq should be left to the president and his advisers?"
In other words: Should Congress propose a timetable, or just leave it all up to Bush?
Here's the Post question, with my emphasis: "Some people say the Bush administration should set a deadline for withdrawing U.S. military forces from Iraq in order to avoid further casualties . Others say knowing when the U.S. would pull out would only encourage the anti-government insurgents . Do you yourself think the United States should or should not set a deadline for withdrawing U.S. forces from Iraq?"
That's awfully close to: Are you in favor of cutting and running? What's amazing is that 47 percent of Americans said yes.
It's fairly easy to manufacture public opinion by careful wording of the questions. Or less careful, too. Have you finally started flossing, by the way? Answer yes or no.