SUSA is full of shit

[thewiseguy]

How could Clinton lead Obama by 18 points in Pennsylvania? :eyes:

They are the only poll that even have the margin in double digits let alone 18 points.

I am not so sure SUSA is the great poll everyone thinks them to be. Here is what they released on their final Ohio poll.

In greater Cincinnati, Clinton had trailed in two previous SurveyUSA tracking polls, but today leads by 19 points.

Obama took Cincinnati 61-38 on the election day! So basically SUSA was off by 42 points in Cincinnati on their final poll.

I believe they have just been somwhat lucky this year with their predicted margins.

[democrattotheend]

1. Their subsets are sometimes off, but their overall numbers tend to be dead on

Their regional or racial breakdowns rely on a smaller subset, so there is a higher margin of error. That is true of any poll.

[Bullet1987]

2. There's a myth here on DU that they've been the most reliable the entire

primary season which isn't true. They've been pretty reliable ever since Super Tuesday, before that Zogby was the reliable one. I don't understand this polling bias seen on this board...no poll will ALWAYS be right.

[Alhena]

3. SUSA shows Hillary winning SE PA, which includes Philly

every other poll shows Obama winning there- poor polling in that area explains SUSA's bad result.

[MadBadger]

4. Greater Cincinatti is not the same thing as just Cincinatti

And Actually they were dead on in their pre-Ohio poll, giving her a 10 point victory.

[thewiseguy]

5. The results I posted were from Hamilton county

They were very wrong on their Cincinnati prediction

[MadBadger]

6. But they were right in their overall prediction, and isnt that what matters?

BTW, i think they are wrong on this one.

[thewiseguy]

7. No, because their Cincinnati prediction did not change the overall results much

Now imagine if they made the same mistake in Philly and Pittsburgh!

[Zynx]

13. Add up the votes from the surrounding counties as well to get their definition.

Also, that's a very small subset of that overall poll.

[rocktivity]

8. GET THEE BEYOND ME, POLLS!

UNCLEAN! UNCLEAN!!!


x(
rocknation

[Political Heretic]

9. Dunno, but that's why I used Real Clear Politics - averages several polls together.

INCLUDING SurveyUSA, Average is Clinton +6.7%
http://www.realclearpolitics.com/epolls/2008/president/democratic_primaries.html

Pennsylvania Democratic Primary
Tuesday, April 22 | Delegates at Stake: 188
Poll.............Date.....Sample......Clinton...Obama....Spread
RCP Average......04/02 to 04/08..-......48.7.....42.0....Clinton +6.7

PPP (D) 04/07 - 04/08 1124 LV 46 43 Clinton +3.0
Rasmussen 04/07 - 04/07 695 LV 48 43 Clinton +5.0
SurveyUSA 04/05 - 04/07 597 LV 56 38 Clinton +18.0
Strategic Vision (R) 04/04 - 04/06 LV 47 42 Clinton +5.0
Quinnipiac 04/03 - 04/06 1,340 LV 50 44 Clinton +6.0
Insider Advantage 04/02 - 04/02 659 LV 45 42 Clinton +3.0

[nebula]

10. You can make a pollster say anything you want them to

with the right amount of cash money.

[L0oniX]

11. They are polling from a list that the RNC gave them?

[Robbins]

12. Missouri

They claimed Hillary was 12 points ahead and Obama won by 1.So that Is a 13 point off.I am a little
skeptical every other poll has Obama gaining In PA and they claim Hillary has risen.

[AnarchoFreeThinker]

14. see....

http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=132&topic_id=5436920&mesg_id=5436920

[Eurobabe]

15. Read this on SUSA from Pollster.com

Very interesting, re: their polling methodology. From 2/08

http://www.pollster.com/blogs/surveyusa_report_cards_a_corre.php

[OzarkDem]

16. ..

:tinfoilhat: :rofl:

[AX10]

17. Here goes the left wing again.

[Texas Hill Country]

18. the poll variations are crazy, but it is an older state and ....

what is interesting is that when you look at the breakdowns of all 4 polls that have a range of Clinton + 19 to Obama + 2, the "white" demographic is consistent in every single one... Hillary 48-52 and Obama 32-37...


and since Pennsylvania has an overwhelmingly white demographic (not to mention it has the oldest average age of any state), I would be suspicious about the polls that have them much closer than 10pts

[4_Legs_Good]

19. 18 points is really nothing surprising. Yes, other polls are showing Obama closing, but

I really don't expect him to actually close Pennsylvania. It's a state tailor made for Clinton. I expect a big Clinton victory.

Don't count on SUSA, but don't count them out. Remeber how utterly wrong Zogby's been all along. I FULLY expect a poll from Zogby maybe late next week saying that Obama is ahead by 15 points in PA.

The on the ground analysis I've seen looks bad for Obama in PA, and has for some time. It's a small roadblock, but nothing devistating. She may pick up 20 delegates, but it won't help her much in the end.

Remember Clinton's BIG DAY last month, where we saw her BIG VICTORIES? In the end, she netted SIX delegates. Big freaking whoop. Even a HUGE victory which gives her 20-30 delegates won't do anything to move her campaign. I'm just warning about putting too much emphasis on Obama winning a knock out blow in PA. Almost certainly won't happen.

Just like his likely big victory in NC won't seal the deal. I'm in it for the long haul, though I am kinda getting tired of the whole thing.

David

[CreekDog]

20. It's not the poll but the turnout model that is critical

each poll has it's own method of weighting the results they get from their phone calls and SUSA's model has differed from many other polls.

[scheming daemons]

21. I think this poll is accurate....

...

and that Hillary will win PA by 18 points.


I'm just hoping that we somehow get a sea change of momentum and lose by only 10 or so.


If we manage to lose by single-digits... why, Hillary would have to quit the race with such a disappointing showing when compared to the SUSA polling.



;-)