KFOR-SurveyUSA Oklahoma poll: Links to crosstabs
By now you've heard about the KFOR-SurveyUSA poll of 652 likely Oklahoma voters from September 5-7. The poll shows Republicans with substantial leads in statewide races:
President: McCain/Palin over Obama/Biden, 65% to 32%.
Senator: Jim Inhofe over Andrew Rice and Stephen Wallace, 56% to 34% to 6%.
Corporation Commissioner (long term): Jeff Cloud over Charles Gray, 52% to 34%.
Corporation Commissioner (short term): Dana Murphy over Jim Roth, 54% to 36%.
The links above will take you to the crosstabs for each poll, showing how the candidates when the sample is broken down by race, gender, party affiliation, age, education, ideology, church attendance, income, and abortion views.
Keeping in mind that the smaller the subsample, the bigger the margin of error, it's still striking that McCain has the support of 42% of Oklahoma Democrats.
I'm happy to see my friend Dana Murphy doing so well. She is the most qualified candidate for Corporation Commissioner that I have ever seen on the ballot. If Oklahoma voters can look beyond party affiliation, she ought to win by a landslide.
0 TrackBacks
Listed below are links to blogs that reference this entry: KFOR-SurveyUSA Oklahoma poll: Links to crosstabs.
TrackBack URL for this entry: https://www.batesline.com/cgi-bin/mt/mt-tb.cgi/4405
I find it ironic that the Democrats perform worse in the Eastern part of the state (the more Democratic-half)...
You are quoting a survey done by a very partisan firm.
I think all four races quoted overstate the republican advantage.
I will bet you that the democrat gets a higher percentage than that in all four races. The democrat has to a higher percentage in all four or you win.
Wanna bet lunch?
No. Because of the number of undecideds, both sides will probably wind up with a higher percentage than the poll shows. In the Senate race, the people currently backing the independent may very well break toward Rice.
SurveyUSA was founded in 1990 by Jay Leve, a former Miami Herald reporter. As far as I know, SurveyUSA does work for TV stations and newspapers, but does not work for campaigns or parties. I appreciate the fact that they publish all the crosstabs -- not only percentages, but raw numbers -- which allows anyone to look at the data and evaluate it.
They nailed the result in the presidential primaries. In 2004 in Oklahoma, they hit Kerry's percentage on the nose and slightly underestimated Bush's percentage. In the Senate race they had Coburn 47, Carson 39, Bilyeu 8. The actual result was 53-41-6. Just before the 2004 Senate primary, they had Coburn 54, Humphries 25, Anthony 13; the result was 61-25-12. Their median bias in 2004 -- measuring results versus their final poll -- was 0.2 points in the direction of the Democrats.
There aren't too many polling firms that put all their data out there and leave it out there for years, for the general public, without charging a fee for access. If you want to prove SurveyUSA bias, you have plenty of data to work with.
It seems that, for Federal Gov. races, Republican campaign ads should only need four words: Harry Reid Nancy Pelosi. What stronger argument is there?