Beware of the Newsweek Poll (CRUSH KERRY)

[sitetest]

You're welcome.

[hawaiian]

GUYS, the poll IS as good as it seems. Newsweek already adjusted for the skew. Wait 'til Gallup's comes out, which will be all post-RNC responses (the 54-38 kind).

[NYC Republican]

If you used more realistic figures, like D-38%/R-35%/I-27% (which was about the ratio voting in the '00 election), and use the same percentages, the vote tally would be:

Bush/Cheney - 50%

Skerry/BreckGirlSilkPony - 43%

Still a nice bump, but not the amount being reported by Newsweek.

[sitetest]

Dear NYC Republican,

"If you used more realistic figures, like D-38%/R-35%/I-27% (which was about the ratio voting in the '00 election), and use the same percentages,..."

That assumes two things:

1. The ratios haven't shifted in four years;

2. Folks self-identification by party label is highly stable.

I don't know that either assumption is valid. I'd especially question #2, especially right after a successful Republican convention. I wouldn't be surprised if some folks who normally identify as Dems will identify either as Reps or as Inds, at least for a few days. This year, that might be even more pronounced with Mr. Miller's highly-successful speech elucidating why someone who once considered himself a mainstream Dem might now think the party has left him.

I think Newsweek built in some weighting (otherwise the overall result would have been 54% - 40% rather than 52% - 41%), but I suspect they may have used other demographic variables more than party affiliation.


sitetest

[tiki]

Never. 4 more years, 4 more years....

[TheLawyerFormerlyKnownAsAl]

"I don't think "weighted" is the right word. That implies intent. It may be that more people id with Rep. now."

If there are more republicans than democrats, which I think is the case, shouldn't a realistic sample have more republicans than democrats? Not a
close follower of polls, but just wondering.

[Bonaventure]

According to Mickey Kaus, the Time Magazine poll has almost exactly the same weighting as the Newsweek Poll. Kaus quotes the Minuteman blog, which speculates that both polls oversampled Republicans because Republicans were home during the convention to take phone calls, while Democrats were out doing other things.

[Future Useless Eater]

>>According to Mickey Kaus, the Time Magazine poll has almost exactly the same weighting as the Newsweek Poll. Kaus quotes the Minuteman blog, which speculates that both polls oversampled Republicans because Republicans were home during the convention to take phone calls, while Democrats were out doing other things.

There was a similar apparent oversampling in the Democrats
favor after the DNC and we howled to no avail.

Maybe its just a factor of people being charged up by the
convention, not really knowing how they are registered,
and more of them CLAIMING to be Democrat after a DNC,
or Claiming to be Repub after a RNC.
We're only talking 5 or 6 percent.

I really don't know how I'm registered, its been so long,
and I wouldn't be surprised if 5 or 6 percent were in the
same boat.