Eugene Volokh has spent a lot of time debunking surveys that purport to be meaningful but aren't. (See also here, here, and here, among others.) Eugene has identified many cases where a self-selected, nonrandom sample is used -- but I'll bet he's never come across a poll which uses Tom Friedman's trick: simply making up stuff. It's breathtaking in its brazenness:
Recently, I've had the chance to travel around the country and do some call-in radio shows, during which the question of Iraq has come up often. And here's what I can report from a totally unscientific sample: Don't believe the polls that a majority of Americans favor a military strike against Iraq. It's just not true.Ah. Well, Gallup could certainly save time if they used the "Follow Tom Friedman around the country" methodology.
That's one thing I love about the New York Times -- they take their role as "opinionmakers" seriously. Not only do they tell us what we should think, but they tell us what we do think.