IN SurveyUSA: Hillary leading by 12

[ChrisFromNJ]

http://www.surveyusa.com/client/PollReport.aspx?g=777e9395-9b22-44ec-a525-981fcb9029e9

5/5/08

Hillary: 54
Obama: 42

24 hours till votes are counted in the Indiana Democratic Primary, Hillary Clinton defeats Barack Obama in the symbolically important popular vote, and possibly by enough to pick up more than a trivial number of net Convention delegates, according to SurveyUSA's final pre-primary tracking poll conducted for WCPO-TV Cincinnati and WHAS-TV Louisville. In 4 tracking polls over the past 5 weeks, Clinton has never polled lower than 52%, Obama has never polled higher than 43%. At the wire, they finish: Clinton 54%, Obama 42%. Among males, the two have been tied in 3 of the 4 tracking polls. Among females, Clinton has always led by at least 14, and finishes ahead by 22. Among Republicans and Independents, the two are effectively tied. Among Democrats, Clinton finishes ahead by 19. Clinton leads among Conservatives, Moderates and Liberals. She leads among Pro Life and Pro Choice voters, among regular and not-so-regular church goers. In Northern Indiana, she leads by 11. In Central and Southern Indiana, she leads by 27. In greater Indianapolis, Obama leads. Among voters under 35, Obama leads. Among voters over 35, Clinton leads.

Ouch for Obama.

[So the Heroes Fall]

Hillary is not going to win 21% of blacks though we can chalk that up to subsample MoE. I highly doubt that blacks will make up only one percentage point more of the electorate than they make up of the state at large though.

[ChrisFromNJ]

So now  the Obamites are trying to discredit the most reliable pollster of this primary season because it doesn't fit their hypothesis of St. Obama's performance in Indiana. Nice.

Yet if Obama was up 12 in Indiana, you'd all be extolling the virtues of SurveyUSA's terrific track record.

[Aizen]

arg, I wanted Obama to win here. Now he's going to fail to seal the deal. Again.

[So the Heroes Fall]

arg, I wanted Obama to win here. Now he's going to fail to seal the deal. Again.

I never expected Obama to win here and no one should've with the demographics. I care more about holding Hillary to only a net 5 or so delegate gain.

So now  the Obamites are trying to discredit the most reliable pollster of this primary season because it doesn't fit their hypothesis of St. Obama's performance in Indiana. Nice.

Yet if Obama was up 12 in Indiana, you'd all be extolling the virtues of SurveyUSA's terrific track record.

*sigh* Can you at least argue with the numbers rather than just use ad hominems?

I don't think the poll is THAT far off but one can hardly argue the internals look right. And if you think it's not setting up to my expectations read what I said above.

[ChrisFromNJ]

*sigh* Can you at least argue with the numbers rather than just use ad hominems?

I don't think the poll is THAT far off but one can hardly argue the internals look right. And if you think it's not setting up to my expectations read what I said above.

The internals look just fine. Hillary seems to be picking up a bit of black support and white numbers look just right. Even if Obama gets 90% of the black vote, he will still lose by close to 10 points. What is your point? What is so wrong about these numbers that you feel the need to discredit the most accurate pollster? You'd be extolling the virtues of Survey USA if they reported Obama leading by 12...

[So the Heroes Fall]

The internals look just fine. Hillary seems to be picking up a bit of black support

Do you have any proof of this other than the poll? I even gave the reason for that, the black subsample has a size of around 67. Thus a margin of error of around 12 points.

and white numbers look just right. Even if Obama gets 90% of the black vote, he will still lose by close to 10 points.

Yes. Not too far off from what I expect and within the MoE actually, but I think common sense would say that blacks would make more than 10% of voters in a Democratic primary in a state that's 9% black.

[ChrisFromNJ]

Do you have any proof of this other than the poll? I even gave the reason for that, the black subsample has a size of around 67. Thus a margin of error of around 12 points.

My proof is in this poll, from the most reliable pollster of the primary season. Your "proof" is your own wishful thinking.

[So the Heroes Fall]

Do you have any proof of this other than the poll? I even gave the reason for that, the black subsample has a size of around 67. Thus a margin of error of around 12 points.

My proof is in this poll, from the most reliable pollster of the primary season.

With a sample of 67 and a margin of error of 12. Jesus Christ.

Your "proof" is your own wishful thinking.

No, it's math.

[Senator Sbane]

I gotta agree with BRTD here the black sampling seems off. I think there are going to be more blacks, around 12%, and I do not see why they would not vote 90-10 for Obama. If you correct that Hillary's margin falls to around 8 points, but which is still an impressive victory for her.

[So the Heroes Fall]

but which is still an impressive victory for her.

Considering the state's demographics, not really.

[ChrisFromNJ]

but which is still an impressive victory for her.

Considering the state's demographics, not really.

So Obama has an excuse in states with a large amount of whites?

What will his excuse be in the general election? Will he write off whites in the general election?

He won't be able to get away with 40% of the each state's electorate being black in the general election.

This would be an impressive win for Hillary, and to say otherwise is hackery at its finest.

[So the Heroes Fall]

but which is still an impressive victory for her.

Considering the state's demographics, not really.

So Obama has an excuse in states with a large amount of whites?

What will his excuse be in the general election? Will he write off whites in the general election?

He won't be able to get away with 40% of the each state's electorate being black in the general election.

This would be an impressive win for Hillary, and to say otherwise is hackery at its finest.

Indiana is even worse demographically for Obama than Pennsylvania and Ohio, so how is underperforming those states if she gets only 8-9 points impressive?

And Obama can write off Indiana in the general election, because he'll never win it in a million years (and neither would Hillary.) Obama will however win my lily-white state, and he would've won it in a primary too.

[Prez Duke]

HAHA. We now have no accurate pollsters in existence. NC is too close so everyone is off and Hillary is winning by too much in Indiana for it to be accurate.

I have a question. Which polls will you actually trust? You all point to the southern states before 2/5 as why they are all wrong but never give them credit for being in the ballpark. If polls show NC CONSISTENTLY within 10%, it's silly to say Obama will still win by 20% there.

[Grad Students are the Worst]

So Obama has an excuse in states with a large amount of whites?

Nope.  But in a 50/50 national race between Clinton and Obama, Hillary getting low double digits in Indiana (a heavily WASP-y, but working-class state) is not impressive.

What will his excuse be in the general election? Will he write off whites in the general election?

He won't be able to get away with 40% of the each state's electorate being black in the general election.

Like Clinton wouldn't win blacks heavily in the General?  Please.

A while back, when Obama was solidly leading national polls and Clinton wasn't, he was still behind among Democrats.  I'm disappointing to see you echoing this "Obama as Mondale" line.  Primary bases are just not representative of General Election bases, although they may correlate.

Clinton will be heavily reliant on blacks.

Obama will be heavily reliant on the working-class.

This would be an impressive win for Hillary, and to say otherwise is hackery at its finest.

This is not the kind of impressiveness that Clinton needs to pull off this nomination, so it doesn't much matter.

[Grad Students are the Worst]

HAHA. We now have no accurate pollsters in existence. NC is too close so everyone is off and Hillary is winning by too much in Indiana for it to be accurate.

I have a question. Which polls will you actually trust? You all point to the southern states before 2/5 as why they are all wrong but never give them credit for being in the ballpark. If polls show NC CONSISTENTLY within 10%, it's silly to say Obama will still win by 20% there.

Are you talking to me?    Because I haven't seen anyone arguing those points here, and BRTD hasn't even spoken to the quality of the poll besides the high-MoE nature of the black sub-sample.  Everything he has said is correct.

[Senator Sbane]

What I meant Hillary would have an impressive victory was an impressive victory in my books. If  Obama could muster up around 42-43% of the white vote I would have been fine with that. But it seems like southern Indiana will crush mine and the black man's hopes. I will be more interested in the rural northern counties though. If Obama is getting crushed there by 20 points, he is in trouble in the general. Of course none of this really matters unless Hillary wins by 15 points or so and that will not happen.

[I left.]

So now  the Obamites are trying to discredit the most reliable pollster of this primary season because it doesn't fit their hypothesis of St. Obama's performance in Indiana. Nice.

Yet if Obama was up 12 in Indiana, you'd all be extolling the virtues of SurveyUSA's terrific track record.

You're becoming an extremely annoying hack. Learn to accept reasonable arguments and at least consider their merit before resorting to cheap taunting.

[ChrisFromNJ]

So now  the Obamites are trying to discredit the most reliable pollster of this primary season because it doesn't fit their hypothesis of St. Obama's performance in Indiana. Nice.

Yet if Obama was up 12 in Indiana, you'd all be extolling the virtues of SurveyUSA's terrific track record.

You're becoming an extremely annoying hack. Learn to accept reasonable arguments and at least consider their merit before resorting to cheap taunting.

When I finally see a reasonable argument from an Obamite, I'll accept it. I haven't seen one yet, just more wishful thinking and more nitpicking of a pollster that has had a terrific track record and just produced a poll that didn't show St. Obama leading by 30 points.

[Tender Branson]

What´s up with SurveyUSA's last North Carolina poll, as well as PPP's Indiana poll ?

[Joe Republic]

So now  the Obamites are trying to discredit the most reliable pollster of this primary season because it doesn't fit their hypothesis of St. Obama's performance in Indiana. Nice.

Yet if Obama was up 12 in Indiana, you'd all be extolling the virtues of SurveyUSA's terrific track record.

You're becoming an extremely annoying hack. Learn to accept reasonable arguments and at least consider their merit before resorting to cheap taunting.

When I finally see a reasonable argument from an Obamite, I'll accept it. I haven't seen one yet, just more wishful thinking and more nitpicking of a pollster that has had a terrific track record and just produced a poll that didn't show St. Obama leading by 30 points.

You're certainly welcome to highlight discrepancies in the statistical sub-samples of any poll that shows Hillary Clinton doing better than expected, just as two supporters of Barack Obama have done here.  I should hope that your belief that such criticism is "nitpicking" was merely hyperbole.

[I left.]

So now  the Obamites are trying to discredit the most reliable pollster of this primary season because it doesn't fit their hypothesis of St. Obama's performance in Indiana. Nice.

Yet if Obama was up 12 in Indiana, you'd all be extolling the virtues of SurveyUSA's terrific track record.

You're becoming an extremely annoying hack. Learn to accept reasonable arguments and at least consider their merit before resorting to cheap taunting.

When I finally see a reasonable argument from an Obamite, I'll accept it. I haven't seen one yet, just more wishful thinking and more nitpicking of a pollster that has had a terrific track record and just produced a poll that didn't show St. Obama leading by 30 points.

then you're guilty of the same selective reasoning and wishful thinking that you're accusing Obama supporters of. Or you're just kind of thick.

[Prez Duke]

HAHA. We now have no accurate pollsters in existence. NC is too close so everyone is off and Hillary is winning by too much in Indiana for it to be accurate.

I have a question. Which polls will you actually trust? You all point to the southern states before 2/5 as why they are all wrong but never give them credit for being in the ballpark. If polls show NC CONSISTENTLY within 10%, it's silly to say Obama will still win by 20% there.

Are you talking to me?    Because I haven't seen anyone arguing those points here, and BRTD hasn't even spoken to the quality of the poll besides the high-MoE nature of the black sub-sample.  Everything he has said is correct.

No. If I was talking to you, you would know it. I am addressing people in general when it comes to all these polls. We are tossing out all the ones that show Obama further behind than we want. NC isn't going to poll like Alabama or Georgia or even Virginia. VA was when Obamamania was at its highest and there are less white liberals in NC by far than in VA. It will not be a 20% blow out in NC, period.

[So the Heroes Fall]

To further explain things, while SUSA's subsamples are dubious for the reason listed, they might be helpful in determining on which side of the MoE the poll falls. In this case it's clear that the black sample is too pro-Hillary (though within the subsample's very large MoE), and blacks were undersampled. True, if you adjust the black vote you still get Hillary by 8-9 points, within the MoE, but that's exactly my point. I'm not arguing the poll is "wrong" by any chance, just that it's not nailing the exact result, something that no one should expect.

[Sam Spade]

If one out of every three voters identifies as a Republican or Independent, it is quite possible the black % of the vote could be as low as 10%.