Charles Foster Kane defeated

FRAUD AT POLLS!

Today's champion in the media framing freestyle goes to -- aw, you peeked.

Once again, a fairly competent piece of survey research is frog-marched to the top of the page and tortured by the cousins at the Fair 'n' Balanced Network. Again, don't blame the "pollsters"; my only major complaint with this lot is that they only stay in the field two days (two evenings, in this case). Well, and there are those annoying only-at-Fox questions that get thrown in at the end:

As far as you know, are Mormons Christians or not?

... but at least it can't poison the rest of the survey from the end.

One way of assessing how much bias is inherent in the instrument is to look at the questions themselves (and at interviewer instructions, like whether the order of names is rotated). Note that when the Opinion Dynamics folks ask about what respondents like in their candidates, the question is open-ended. Any distortion from this one is down to Fox, which has a very Sam-and-Gollum attitude toward the numbers: Once you give me a coney, the coney's mine, see, to cook, if I have a mind. That sort of takes us back to the days when press barons stalked the earth, which isn't necessarily a Good Thing.

None of which addresses the question of whether polls are ever news. I still think they are -- just that there are almost no circumstances under which they're anything resembling frontpage news. What makes them valid is rarely heddable, and what makes them heddable is rarely valid.