Posted on 09/04/2004 5:49:17 PM PDT by American100
Did anyone else see this today? Looks like the liberal media is at it again.
Beware the Newsweek Poll: The latest poll from Newsweek has the President leading by 11. But don't get too excited. Their turnout model is weighted 37%-31% in favor of Republicans. Though it would be great, it bears no resemblance to history. We've warned you about some of these polls before. Don't get lulled into a false sense of security. Keep working.
from www.crushkerry.com
37 % I had 7% heavy on Republicans in wich case that is a huge skew!
How do you explain away the Times Poll though?
As for NewsWeek all there polls are junk -
Also, the real numbers might be a bit higher anyway since one of the polling days was before Bushes Speech. After Bushes Speech he was winning by 16.
Post-convention polls traditionally lean in the favor of the party who just had the convention.
I wouldn't be suprised if Arnold shifted 5% on his own "You are a Republican if...." speech.
I would like to see more national polls on this, but there is no reason to believe that Bush isn't up ATLEAST 8-10 right now.
I don't know the internal specific of the Time poll. The numbers are looking good - I remember back in 2000 Newsweek gave Al Gore a 10 point bounce out of his convention. It's going to be a close race and we still need to work as hard as we can.
Dear American100,
Although 38% of their sample was Republican, and 31% of their sample was Democrat, I think it's likely they weighted Republicans and Democrats differently in coming to their final numbers. Here is a link to a post I made on another thread examining this question:
http://www.freerepublic.com/focus/f-news/1207324/posts?page=255#255
The raw numbers would have gone 54% for Bush, 40% for Kerry if they'd have computed the poll without doing any weighting.
sitetest
Fair point.
Keep your powder dry, although I do believe Kerry will be flattened like Dukakis.
Excellent analysis.
It jives with Bush being up 16 after his speech.
Also only 12% minority I read.
No matter how the sample ratio has been modified (turnout %, etc.), it is difficult to imagine a national sample with a 7% greater representation by Republicans. It is likewise difficult to believe that the Democrat and Independant samples are equivalent. IMO the pollsters are busy trying to favor the Republicans by 10%+ in the sample so that they can run a "corrected" poll in a few weeks showing a "precipitous decline" in GW's numbers following recent "events and disclosures" forthcoming by the Dimo establishment.
The only polls I semi-trust are Strategic Vision polls.
Dear Mad_Tom_Rackham,
As I pointed out in #8, it doesn't appear that Newsweek used a weighting of 38% R, 31% D, 31% I. That's just the raw numbers.
sitetest
I don't think "weighted" is the right word. That implies intent. It may be that more people id with Rep. now.
Question.
Am I correct in thinking that the 38%R, 31%D, and 31%I are just the breakdown of the folks first reached on the polling call?
And that the pollsters compared these numbers to the expected turnout numbers and weighted them to fit the turnout model before releasing them?
The great fun in the pole numbers is the effect it's having on the DU.
Dear EllaMinnow,
Answer:
I don't really know for sure.
What I know is that they're saying that the respondents to the poll were 38% R, 31% D, and 31% I.
Then there is this:
Bush receives 94% of Rs, 14% of Ds, and 45% of Is.
Kerry receives 4% of Rs, 84% of Ds, and 40% of Is.
(This data is taken from the internals reported by Newsweek.)
If you do the arithmetic, it looks like this:
For Mr. Bush:
R 94% of 38% = .3572
D 14% of 31% = .0434
I 45% of 31% = .1395
.3572 + .0434 + .1395 = .5401, or 54.01%
For Mr. Kerry,
R 04% of 38% = .0152
D 82% of 31% = .2542
I 40% of 27% = .124
.0152 + .2542 + .124 = .3934, or 39.34%
Thus, the raw results, which can also be derived from the number of interviewees for each candidate, come to about:
Bush: 54%
Kerry: 40%
But Newsweek reports 52% - 41%.
That suggests to me that they weighted the respondents in a different proportion than that which they got from the raw numbers. In other words, they counted the Dems more heavily than the Reps, to make up for the discrepancy in their raw numbers.
sitetest
Your theory make sense to me. Thank you for posting your analysis.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.