When Polls, and Pundits, Deceive

The moment I read about this Quinnipiac University poll, I knew it was trouble.

Supposedly, the data show that having the endorsement of a gay-rights group is a handicap for a national political candidate. Objectively, what they show is that, in the three large swing states states surveyed, somewhat more respondents said they'd be less likely to vote for a candidate who was endorsed by a gay-rights group than said they'd be more likely—although in all three states, an outright majority said it would make no difference.

Now, any pollster knows there's a huge difference between what people say they'd do in a hypothetical situation and what they actually do in real life. Here's an analogy: In 2005, there were 539 murders in New York City. Does that make you more or less likely to move to New York? In real life, it probably doesn't make a difference; you're moving for a job or to be closer to your family, and if you do worry about crime, maybe you know that, per capita, New York actually has a lower murder rate than most other big cities. But in the context of a survey, forced to react to that one, isolated piece of information, you're probably going to say "less likely."

Just to be sure I wasn't making anything up, I ran this line of reasoning past a leading pollster, John Zogby. He agreed that the number of single-issue homophobes voting in the 2008 presidential election is likely to be relatively low: "There is a war going on."

So what purpose does a poll like Quinnipiac's serve? Apparently, to give fodder to "traditional-values" demagogues like Bill O'Reilly—who still had to utterly misrepresent the findings to make their point.

(YouTube clip via Newshounds. Listen to the utter contempt in O'Reilly's voice when he says Giuliani has been "progressive" about gay rights around the 3:30 mark.)