Obama 49, Clinton 39
Yikes, two SUSA polls out today that are wildly different from other polling firms. I think I'm apt to trust the PPP polls since they are actually based IN North Carolina. They have also adapated their "Likely Voter" window, which I don't think SUSA has done.
Both PA and NC are big states that may prove difficult to poll. Maybe SUSA is off here? Just like any polling firm they are susceptible to polling errors?
Per my comment on the PA poll post: Just because SUSA got CA right and they were leading in a report card that THEY PUT TOGETHER, doesn't mean we should just blindly trust them (they were off by 12 in MO). We should interpret their results like those from ANY pollster.
Posted on April 8, 2008 5:52 PM
In the second paragraph it says quite clearly SurveyUSA determined some 700 contacts to be likely voters, and that those are the ones that comprise their results.
Posted on April 8, 2008 6:14 PM
I understand that. Let me clarify: PPP "changed" or "updated" their Likely Voter window to better account for the types of voters that have been voting in previous contests. Maybe SUSA hasn't changed theirs since 2006 or 2004?
Posted on April 8, 2008 7:51 PM
I guess the context for judging NC is that a bunch of similar states, regionally and demographically, have voted, from VA to GA, and Obama has won all of them by well over 10 points. And Obama is obviously not suffering a national decline, as measured by the Gallup and Rasmussen tracking polls.
So it will take more than one poll for me to be convinced that he's only up 10 in NC. And we will have plenty more polls before the primary.
Posted on April 8, 2008 8:29 PM
I think I'm apt to trust the PPP polls since they are actually based IN North Carolina.
That is the most simplistic analysis I've seen her in a long time. Can you prove that PPP has been more accurate than Survey USA in NC polling in the past?
Just because one pollster is based somewhere, doesn't mean pollsters from other states will have trouble gathering data on the racial, age, economic (etc.) composition of that area.
Now unless you have data from 2004 and 2000 showing that PPP is more accurate than anyone else in NC, or that the Field Poll has been more accurate than anyone else in CA, etc., your theory is shaky.
Posted on April 8, 2008 8:32 PM
Kingsbridge...Mike didn't say that he thought that PPP was more accurate than SUSA in the past. What he said was that they were more likely to modify their poll based on local demographic trends in "Likely Voters"..
In fact ALL of the pollsters underestimated Obamas victories in S. Carolina, Georgia, and Virginia.In S.Carolina SUSA predicted a Obama 43% Clinton 30% Edwards 22% [a 13% differential] Result. In fact Obama polled 55% to Clintons 27% and Edwards 18%...a 28% differential. The closest poll, PPP predicted a 20% Obama-Clinton differential...still a major underestimate (wrong by 8%). The worst was ARG that said there would be a 3% differential. Reuters was off by 13%, Mason-Dixon by nearly 20%!!! And it should be noted that SUSA had it a 36%-46% Obama lead just a week before the election.
SUSA didn't participate in the Georgia surveying. But much the same thing happened. Obama defeated Clinton 66%-31% (a 35% differential). Reuters/Zogby came closest...with a Obama 49%-Clinton 29% poll (20% differential). Again McClatchy/Mason-Dixon stumbled...with a poll suggesting a close race Obama 47%-Clinton 41% It makes you wonder who they were interviewing?). PPP, Insider Advantage, and Rassmussen all had results suggesting Clinton would hold the race to within 15%. They were off by 20%!!!
They were closer in Virginia, where Obama won 64% to 35% (+29%). The best poll here was actually SUSA who had a survey with Clinton 38% vs. Obama 60% (+22% differential). All the others thought the race would be within 15%-20%...as did a SUSA survey just three days before the election.
SUSA and the other pollsters also vastly underestimated Obamas win in Alabama, where he won by 16%. They all had polls showing it to be a statistical tie.
Posted on April 8, 2008 9:37 PM
Cinnamonape, that's not the part of Mike's analysis that I criticized. I'm aware of his argument about the modification of "local democraphic trends". But I was talking about something else. I have a problem with one of his premises, where he explicitly said that because PPP is based in NC, then it's more "apt" than Survey USA.
Posted on April 9, 2008 7:19 AM
Comments: (you may use HTML tags for style)
Please be patient while your comment posts - sometimes it takes a minute or two. To check your comment, please wait 60 seconds and click your browser's refresh button. Note that comments with three or more hyperlinks will be held for approval.
Please email us to report offensive comments.
See our comment policy here. Note that we require commenters to share their email address via Typekey. We will never share your email address with anyone without your explicit permission.
MAP - US, AL, AK, AZ, AR, CA, CO, CT, DE, FL, GA, HI, ID, IL, IN, IA, KS, KY, LA, ME, MD, MA, MI, MN, MS, MO, MT, NE, NV, NH, NJ, NM, NY, NC, ND, OH, OK, OR, PA, RI, SC, SD, TN, TX, UT, VT, VA, WA, WV, WI, WY, PR