If SurveyUSA is consistently correct, it's going to be a very painful night for Obama supporters like myself. Anyone know their track record?

What's to predict? Volatile polls aside, Obama's camp has already conceded defeat before the polls open by explaining that a loss is still a win. Nice comfort for them but just as pathetic and embarrassing as their claim to win Nevada after losing the caucus rigged in their favor! A loss is a loss.

While I am sure their will be some surprises tomorrow, I think that all that will be left is for the 'fat lady to sing.'

I wonder how gracious Obama will be in his defeat?

I've looked at SurveyUSA's results so far:

FL: Over-reported HRC by 2 points, under-reported BHO by 5

SC: Over-reported HRC by 3 points, under-reported BHO by 12

I haven't found any other surveys they did immediately before NH, IA, or NV.

But draw your own conclusions.

PS: Looking at all the results so far, Zogby looks best, with Rasmussen just okay and ARG similarly skewed like SurveyUSA. Too few data points and too much general data noise to draw any real judgments. But it's interesting to apply those "biases" to the California polls, for example...

I think you can tell how good they are by who are their customers. Survey USA's customers are all top notch. Zogby's (Obama's only great numbers)onlyy customer appears to be a combination of reuters/c-span and I'm sure c-span does not have any money and reuters is for sale.

Fairandbalanced, all the polls have been way, way off. I don't buy any of them. SurveyUSA has been one of the worst. The only issue is the trend and the trend shows mo for obama. Is it enough, who knows. However, the clintons have billed themselves as inevitable since day one and have lost alot in the polls. If they don't have a huge, wipe-out win, it is a victory for obama. Odds are they won't have the huge wipe-out win that they have been preaching. We have a horse race, which is good for democracy regardless of the winner. Tomorrow will be an interesting day.

Here's an illustration that SurveyUSA is not to be given credence. In New York, African Americans are 16% of the entire state population excluding Blacks identifying as Hispanics, plus another 2% Black Hispanic. So 18%. Blacks will be far far more than 18% in the Democratic primary (and of course far less in the Republican primary). The SurveyUSA poll is giving Blacks a weight of 21% in the composition of likely voters -- far too low. Black turnout for Obama was big in the earlier states and they voted for him versus her by 75% to 25% or wider. Perhaps SurveyUSA arrives at its 21% estimate on the basis of old exit polls in NY primaries from years ago when Blacks weren't as keen on voting as this year. I'd give Blacks a weight of 32% this year, though SurveyUSA would say I've got no disciplined method for reaching that number and they've got to use a method.

Whoa Eric!

That prediction regarding Clinton's percentage in New York makes sense only if you give all undecideds to Obama! Distributing undecideds proportionately would give Clinton 58-59%, which I think is much more likely.

Here's an illustration that SurveyUSA is not to be given credence. In New York, African Americans are 16% of the entire state population excluding Blacks identifying as Hispanics, plus another 2% Black Hispanic. So 18%. Blacks will be far far more than 18% in the Democratic primary (and of course far less in the Republican primary). The SurveyUSA poll is giving Blacks a weight of 21% in the composition of likely voters -- far too low. Black turnout for Obama was big in the earlier states and they voted for him versus her by 75% to 25% or wider. Perhaps SurveyUSA arrives at its 21% estimate on the basis of old exit polls in NY primaries from years ago when Blacks weren't as keen on voting as this year. I'd give Blacks a weight of 32% this year, though SurveyUSA would say I've got no disciplined method for reaching that number and they've got to use a method.

The SUSA results quoted up-thread are close to Clinton's results within any plausilbe MOE (2 to3 percent), but well under Obama's. I think that makes sense - when I looked at their CA poll earlier, thought they weren't predicting near enough turnout. Not that I took a real scientific approach to it, but they interviewed about 1700 RVs and came up with about 850 LVs. That just seems way short given the turnout so far. Their trends were consistent with all the other polls, but not their percentages.

It's going to be a lot closer than a 12-point spread.

Here's my write-up as an Australian onlooker.

With 15 hours until the first polls close in the States, these are my final predictions. About seven states have no reliable, recent polling data — in these cases I've tried to go on demographics and media/advertising exposure, but my numbers could be, you know, way out.

I've certainly factored in an Obama surge, expecting he will pick up a lot of the undecideds. So if you're looking for a way to contextualise my prediction, I would say it's slightly (but reasonably!) optimistic for Obama.

The two states to watch, I believe, are the ones where I've indicated a dead heat: Missouri and Arizona.

These states will report results well into the day, so if you're looking for early indications, watch the two eastern strongholds for Hillary: New York and New Jersey. If Obama is a couple of points higher than my predictions after at least 30% of the vote is counted in these states, the surge is bigger than I expected, and he might well carry the day in delegates. If several points lower, the day is Hillary's — although how the US media will spin that is anyone's guess. She could win by anything up to 150 delegates and they might call it a photo-finish. Obama only needs one more delegate than Clinton for a famous victory.

Such is life and politics.

Win or lose, Obama has acquitted himself admirably. Burdened by his name, his race, his lack of recognition, his outsider status, he has won friends and fought Hillary Clinton to something like a tie. Three months ago, this would have been unthinkable. If he had another month, he would probably beat Hillary convincingly. No wonder Billy Shaheen called him a drug dealer. The man is dangerous.

Will the Obama surge translate into actual votes? Will the young votes actually turn out? Or will Latinos and women turn out in large numbers for Hillary?

I predict HRC will win New York; New Jersey; narrowly in Massachusetts and Connecticut; Tennessee; Missouri; Oklahoma; Arizona; New Mexico; and finally, she will get a 5-6 point win in California. Although Obama has been surging in California, Hillary's organization there is very strong, and she will prevail.

As to delegate count, I'm predicting about a 75 delegate win for Clinton. Although the media will have a hard time acknowledging it, Hillary will get a bump out of Super Tuesday, particularly in the fund raising area.

With the WingNuts promising to sit out the election if McCain wins and the Obama fans promising to not vote if HRC wins this just might be the sanest election in American history.
Now how do we insure that the Fruit Loops, left and right, keep their promises to the rest of us?