Polls weighted to reflect LIKELY (not past) turnout in TX and OH show Obama landslides.

[AnarchoFreeThinker]

Decision Analyst selects their poll sample based on likely turnout (not turnout from past elections), and has Obama cruising in TX (57-43) and OH (54-46). This updated sampling method might give a better indication of what the actual vote will be, as opposed to old sampling models that have underestimated Obama's support.

http://www.usaelectionpolls.com/2008/texas.html

[loveangelc]

1. they don't say what their methodology is though...so I can't believe it.

[AnarchoFreeThinker]

3. Sorry, don't have much more to go on but the following:

"They say that they carefully selected the demographic compositions used in their polls. Most pollsters thus far have not been fully representing the support Barack Obama gets at the polls because the demographic composition of those going out to the polls tend to be under-represented in traditional polling based on previous turnout."

grain o' salt

[mohc]

6. From their site

"Results based on scientific sampling of registered voters from American Consumer Opinion® (www.acop.com)"

decisionanalyst.com

While I may want to believe the results of this poll, acop.com is a paid online survey site. Unless they have a very large registration base, the sample will probably be too self-selective to be reliable.

[Dems Will Win]

7. I can explain it. I do polls locally

Historic approach will put youth turnout at anywhere from 4% to 11% for a primary, while the current turnout approach will note that youth turnout in the primaries have been from 13% to an astounding 27%. So they weight the poll differently, allowing for far more youth in the final poll result.

Since the young are 65% to 70% for Obama, you get a radically different result. This is why Obama's numbers have been misunderestimated in polls.

Always use the current turnout approach.

Unless you are stupid and like to be wrong all the time.

As someone noted, the totally accurate Des Moines Register poll used a current turnout approach, while the polls that understimated Obama strength did not.

Obama, as of today, would win both Taxas and Ohio by healthy margins, and will clean up on the delegates in TX.

Do you still have a question about why the DA poll is far superior to the old, obsolete polling?

Don't know if I explained it enough for everyone.

I have won 24 out of 29 county-wide races for progressives as a consultant so I know whereof I speak.

[AnarchoFreeThinker]

8. Thanks, DWW.

[loveangelc]

10. wow that's interesting. at the same time, one poster pointed out that its an online survey.

i don't think online polls are that reliable.

[Dems Will Win]

13. Online polls are suspect but if weighted properly tell you something

The only real poll is back to knocking on doors with a clipboard because of the cell phones. So what you do instead of reaching youth with cell phones is you extrapolate and weight the poll so they are represented by the landline youth.

All it takes is a calculator!

[HeraldSquare212]

2. Remeber the Iowa newspaper poll that everyone thought was crazy

in it's estimate of turnout? I would think other polling orgs are using more than historical data to poll at this point.

[Renew Deal]

4. I don't see what they're basing it on.

I get what they're doing, but I think they're basing it on past numbers only. Is Obama's ground game as good in those states? Is he as entrenched?

[Yurem2008]

5. They know something that we don't know

they were the only ones that called Wisconsin a blowout for Barack Obama. Now they are predicting solid wins for Obama in TX and OH

[Dems Will Win]

9. See my post above. They weight for more youth than the historic-method polls

[loveangelc]

11. when did they predict a wisconsin blowout in wisonsin? i think that was another poll...

[yellowcanine]

12. Another thing about Decision Analyst many are missing - it is a marketing research firm.

That is their bread and butter. Marketing research firms have to know how to poll different demographics. Their clients demand it. After all, you don't want to develop a product for the youth market and inadvertently poll boomers. Many here panned DA because they are relatively new to political polling. But political polling is going to be more accurate if the demographics are correct and DA should have an edge in that regard.

[democrattotheend]

14. Does this pollster have a track record?

Or are they brand new. I hope you are right, but hard to trust them yet.

[TexasObserver]

15. You're right. Polls that use models built on last time are missing the boat entirely.

Obama is winning Texas big time, by 8-18 points.

[grantcart]

16. pls note that decision analyst has no previous political polling exp.

They are however expert in finding out if it would preferable to use brown or yellow packaging for whole wheat bread.

Doesn't mean they aren't right it's just that with all of the demographic groups in Texas this would be a hard place to start and get it right the first time.