It speaks well for SurveyUSA.

So she essentially held serve in most states and he overperformed in a few.  Really, if you take Zogby's crazy polls out of the equation the polling was pretty good in most states.  To the extent it was off, it tended to underestimate Obama.  Again, taking Zogby's lunacy out of the equation.

i never like to say this about any poll.

but zogby is crap .

ever since he released a memo attacking the clinton camp i knew he was full of it.

yes survey usa has been the best out there for a while now.

they scored better than the field poll this go around

Your piece "A Tie Goes To Clinton" is linked from Real Clear Politics!

Way to go.

by the way clinton is leading in new mexico with 98% reporting.

wonder when they'll call it for her

I was studying the difference between Zogby and SurveyUSA yesterday and their methods. In short, SurveyUSA has more of what's called internal "validity" to their surveys. Especially the two surveys with identical results a day apart - that's a sure sign of validity. According to their site they also dial randomly both listed and unlisted numbers, something that Zogby might do but doesn't list doing at his site.

I also think that since Zogby is an Obama Superdelegate that lends statistical bias error to his polls, that while not evident, is a possibility. Also, I believe someone here said Zogby polls only in English - a big significance in California.

I love it when polls go wrong. I am one of those that cant wait to see the public lose all confidence in these polls. Just go vote and not trying to play amateur pollster and spend more time analyzing actual politics because most people do not have a clue on how to interpret data. it is not going to kill anyone to wait a few extra days until election day to figure out who won. Leave the detailed attention to polling to the campaign strategists.

Zogby nailing Missouri only goes to show that a broken clock is right twice a day.

I will be waiting for his memo saying how he saw a Clinton uptick in California the last night of his tracking but didn't report it because it was only one night.

So it looks like, contrary to what the media's been saying, Obama actually outperformed expectations in a lot of states. I don't disagree that the media has been more friendly to Obama than Clinton as a rule, but I think the way they've been treating Clinton winning New Jersey and Massachusetts as an upset is ridiculous, given the margins she had a week ago in the polls. I am also curious why people are making a big deal out of Obama losing Massachusetts, where he was down by 30 a couple weeks ago, and keep highlighting the high-profile endorsements he received there as a reason he should have won it. Yet in treating New Jersey like a swing state, they fail to mention that Clinton had endorsements from all of the big name Democrats in the state. Not to mention half the media market.

What happened with the exit polls? The ones that were posted on the front page here yesterday ended up being drastically off in some cases. Obama was said to be +2 in MA and ended up losing by 15 points. Obama was said to be +7 (IIRC) in NJ and lost by 10 points.

Exit polls are supposed to be very, very solid with very small margins of error. The differences in at least these two cases are large. Anyone have any insight on what went wrong in those two places? Small undemographically sound sample sizes?

CNN says the California Latino vote was 29%.  Holy crap.  One of the polls had it at 27%, and everyone thought it was too high.  It was only 16% last time.