Survey USA - New Polling - Missouri Bush 48 Kerry 46, PA Bush 47 Kerry 49, Indiana Bush 60 Kerry 36

SurveyUSA ^ | Sept 9 2004 | SurveyUSA

[okstate]

Missouri - http://www.surveyusa.com/2004_Elections/MO040910presgovsen.pdf

Pennsylvania - http://www.surveyusa.com/2004_Elections/PA040910pressen.pdf

Indiana - http://www.surveyusa.com/2004_Elections/IN040910presgovsen.pdf

Also polling for Kansas and Kentucky. Bush 60 Kerry 35 in Kansas

http://www.surveyusa.com/2004_Elections/KS040910president.pdf

Kentucky Bush 56 Kerry 39

http://www.surveyusa.com/2004_Elections/KY040910pressenamend.pdf

Kit Bond, Specter, Bunning, Bayh, and Brownback all win easily, according to these polls.

[okstate]

The MO poll is an outlier, IMO

[NavVet]

Kerry has pulled ads in MO, that tells you what their own internal polling is saying.

[okstate]

Yep, and Gallup had President Bush up like 14 points in Missouri

[KJacob]

Kerry has pulled ads in MO, that tells you what their own internal polling is saying.

I heard today that he has changed his mind on that. I do not think it will matter.

[steveegg]

SurveyUSA has been consistently pro-Ketchup in both Missouri and Pennsylvania (thanks for the archiving of the polls, Dales).

[Dr Snide]

just a quick note, this survey like many others showing a small Bush lead are not a true random sample. Instead they are a random sample that has been "weighted" in order to "better represent current demographics". In the case of Zogby at least that means using close to 40% Democrat as a baseline. All polls that use a true random sample show the President up by double digits. The consistent trend appears to be that ALL the true random polls show more Republicans than Democrats. It may be possible that are just simply more Republicans than Democrats now.

[okstate]

SurveyUSA does not weight their polls. That's how they got bizarre results in the past, like when they had Kerry up 12 in PA. Or when they had Kerry up only 1 point in California.

Their samples jump all around in party ID.

[Revolutionary]

I believe this and Rasmussen use a 4% Democrat bias (artificially take 4% more Democrats than Republicans). The last major survey showed that Republicans and Democrats tied at 31% each in the electorate. I believe that since then the Republicans probably have picked up another 2%.

I just add 4% to what ever they come up with. It then tends to match the other polling data pretty well.

[okstate]

Polling companies that use weights:

Rasmussen, Zogby, Fox News (I think).

Polling companies that don't:

Time Mag, Newsweek, Gallup, and almost positive that Survey USA doesn't

[Dr Snide]

read the last page of the PDF file, Survey USA does use weights

[Jewels1091]

I don't understand how these polls can be so different from each other! They are so many points apart!

[Dr Snide]

although as a correction it does not specifically say the survey is weighted by political affiliation but does report weighting by demographic information

[okstate]

I don't think it is weighted by political affiliation.

There's just no way... the numbers don't work out on many other older polls they've done.

[comebacknewt]

SUSA certainly shows a different race than Gallup or the AP suggest. In the past, their state polls have been pretty accurate, so I gues it is possible.

As terrible a candidate as F'n is, this is ultimately going to be another close race. I am afraid that hundreds of millions of RAT dollars and an unabashed propoganda effort by the MSM will get Kerry back within striking distance by election day.

[SolidSupplySide]

I don't want to open up .pdf files on my dialup connection. Could someone please provide the facts of these poll such as MOE, polling dates, etc.?

[jmaroneps37]

.. and the major good news for frenchie that has given him back the lead is ????????
Puleeeeeze!