There are two major problems with the Harris poll that found 14 percent of Americans think President Obama may be the antichrist and has some Dems freaking out.
The first problem is question design and survey format. The poll started by telling people "Here are some things people have said about President Obama" and then asked them to agree or disagree with a series of pejorative statements—from "Obama wants the terrorists to win," to "Obama may be the antichrist." Respondents were not given a set of alternative statements to consider. As ABC's award-winning pollster Gary Langer explains, presenting the questions like that practically guarantees a skewed result. "‘Some people have said' is a biasing introductory phrase," he writes. "It imbues the subsequent statements with an air of credibility—particularly when you don't note that others say something else."
Using such biasing phrases, and giving respondents no alternative premises to consider dramatically enhances what statisticians call "acquiescence bias"—a tendency in survey respondents to agree with all the questions asked of them, especially when respondents are in doubt. Acquiescence bias is particularly pronounced in surveys that employ truisms (like, "do unto others as you would have done unto you"). But it also creeps up with inflammatory statements (like "your president is the devil").
The second problem is sample selection. The Harris poll relied on people who signed up to participate in an online survey regarding negative opinions of President Obama in exchange for cash and prizes. A slew of studies have shown that offering people gifts in exchange for filling out an online questionnaire does not produce a random sample—the gold standard in polling. People who opt in to such surveys are a self-selecting bunch and because of that, their opinions and attitudes are not representative of the population at large.
In fact, surveys based on nonrandom samples are so notoriously unreliable—as Langer says, they "lack the theoretical underpinning on which valid research is based"—that the official policy at most news organizations (including The Washington Post and The New York Times) is to ignore them. Not a bad idea.