Poll Results That Mean Exactly the Opposite of What They are Interpreted To Mean

Have you ever had a conversation in which you nodded in agreement with an assertion, but later, when you thought about it you realize that it was totally wrong? Mickey Kaus points out an example:

I never understand poll results like this:

"[L]ess than a third of voters" in three swing states said the Ayers and ACORN issues "would affect their votes."

This is interpreted by TIME, and others, as a sign that McCain's Ayers and ACORN attacks aren't working.

But if the attacks really did move a bit "less than a third"--27% or 30%--of voters into the McCain column, they'd be fantastically effective. ...

Obviously, the attacks haven't done that. Most of the 27%-30% of affected voters were probably McCain voters anyway. But the poll doesn't tell us how many weren't. If even a third of that 27% were undecided voters, that could have a huge impact in a close race.

Nor does the poll tell us how much these voters votes are being affected, or even in which direction the "affect" cuts. (Maybe the attacks produced sympathy for Obama.)

All this is assuming voters are being improbably, brutally self aware and honest about what's affecting them. ....

... They're manufactured news designed to give the impression that lines of attack the MSM doesn't want to work aren't in fact working. ..