Obama 47, Clinton 43... McCain 54, Huckabee 36, Paul 4
Clinton 47, Obama 46... McCain 61, Huckabee 27, Paul 3
Note the Zogby correction. They had inaccurately reported Clinton led by "double digits" in their single day Friday polling. In reality, she was only ahead by a single point. Looks like the red phone ad and negative attacks have done little to move the numbers.
Posted on March 2, 2008 10:32 AM
I do not have access to the detail of today's poll, unlike the one from 2 days ago (2/26-2/28) where Obama was ahead by 6% and was up by 10% among those who already voted. Since early voting is over, what were the results? How does one explain losing a 10% lead in early voting?
Could someone post the demographics used?
For example, respected Survey USA (last poll) used 21%AA, 28%Hispanic, 56%Women, 13%Ind, 9%Rep., 58%under 50, 16% were 65+?? The older Zogby poll had only 5%Rep. and only 48% under 50 and 35% Hispanic, 20%AA.
Therefore if Zogby were to use the demographics for Hispanics, Age and Republicans that SUSA used, Obama's lead would be much bigger.
I suspect Zogby is being extra cautious due to their poll outcome in CA.
The main question I have (being ignored by pollsters and media) is whether Republicans will turn out in much higher numbers and whether they will vote against Hillary - who they despise, or vote for Hillary as recommended by Rush Limbaugh, just to screw up the Democrats?
Posted on March 2, 2008 12:53 PM
Posted on March 2, 2008 12:56 PM
I continue to be surprised by how literally the media takes these polls. Just before logging on to Pollster I saw an AOL headline that read: Obama leads in Texas, tied in Ohio. The story was based on the Zogby poll. But Obama doesn't really lead in Texas, he leads in a poll taken in Texas. The headline also fails to mention that Zogby missed by 23 points in California, 15 points in New Hampshire and 14 in South Carolina. Nor does it say that in a report card of all polls the best ones are getting no closer than 4% to predicting the outcome. In a post on February 5, 2008 (after hearing Tom Brokaw gravelly announce that Clinton was trailing by 13 in Ca.) I cautioned that polls were becoming part of the news story and were starting to alter voter behavior in much the same way as the early release of exit polls. This is a wonderful site and I have learned a tremendous amount from reading Mark's posts and listening to the follow up comments. But the biggest lesson that I have learned is that while polling is a science, it is a very inexact one. The post that Mark made today about Ohio demographics showed how much the numbers can be altered by tweaking the underlying assumptions. It makes me wonder whether polls involving complex issues can ever be predictive. In any case, I think the professional polling companies and the media owe it to the public to be more open about their historical results and the whimsical nature of polling based upon unproven assumptions. I think every poll or media report of a poll should include the statement Mark made this morning. "The tools of survey research are simply not powerful enough to predict who will vote with great precision."
Posted on March 2, 2008 12:59 PM
Posted on March 2, 2008 1:01 PM
A very very good post from richard above. People are starting to live and die by these polls.
Looking at the numbers on RealClearPolitics, which compare final poll averages to actual vote, a trend has emerged: Obama outperforms polls by 5-10 points (on average), whereas Clinton finishes just about where the polls have her. I've been saying this all along:
I think these polls are UNDERCOUNTING the cell-phone, under-30 group, a group that has decisively broke for Obama.
If these latest polls are to follow that trend we could expect a solid win for Obama in Texas and a squeaker in Ohio. Thoughts??
Posted on March 2, 2008 1:07 PM
I always wonder about Zogby polling. Does any one take into consideration that the Zogby brothers are ardent Obama supportors? James Zogby is a Democratic Super Delegate for Obama. Kind of a conflict of interest.
Posted on March 2, 2008 4:03 PM
the conspiracy theories should be left for a different blog. The Zogby who supports Obama is in no way connected to this firm. A lot to do about nothing has been made of the Zogby polls incorrect prediction for CA. EVERY POLLSTER is going to screw up at some point. Anyone who points to that poll is ignoring the tons of other polls that Zogby has nailed. Enough with the conspiracy theories. Can we not just look at the merit of polls without considering whether the pollster is just trying to support a candidate. Not only would that be unethical, but it would be bad science, and I'm guessing no pollster would be in business if he did bad science. Enough of this already.
Posted on March 2, 2008 6:40 PM
So that we can better judge this poll, if someone has access to the demographics, please let us know. However, if they are the same as a few days ago, then Zogby has oversampled the Hispanic vote and undersampled the probable Republican vote in Texas. He has paid attention to who has already voted. Therefore, he is not likely to repeat his mistakes in CA. By tomorrow we should have some more polls. If Obama is ahead in TX, then expect some Major endorsements as an attempt to wind down the election.
Posted on March 2, 2008 8:38 PM
Zogby gave 13 plus points to Obama in California
how can one depend on in Zogby's Ohio/Texas polls. Zogby poll is always tilted towards Obama.
Posted on March 3, 2008 8:36 AM
Zogby gave 13 plus points to Obama in California, how can Zogby polls for Texas and Ohio be releable?
Posted on March 3, 2008 10:04 AM
Every poll is suspect. To paraphrase Lincoln, some pollsters are wrong all of the time, and all posters are wrong some of the time, but all posters are not wrong all of the time.
Posted on March 3, 2008 10:57 AM
Aren't the above points about Clinton finishing where the polls put her and the fact that Obama out-performs in the polls pretty obvious?
In almost every state, Obama, in the last week or two before the election, focuses on a state and overcomes what is usually a 2-digit lead in the polls. This means that Clinton has, in each state, built-in support, as any largely known national figure would.
This means that most of the early voters split for Clinton, and most of the late voters trend towards what the polls show when Obama comes into a state.
Posted on March 3, 2008 10:59 AM
The theory behind polling is an exact science called statistics. To achieve a high confidence level of the polling result, one needs a large, typically 1000, unbiased sample. In reality getting an unbiased sample is difficult and it is not a science anymore. Statisticians can only give the odds and never to predict the outcome. That's why they say statisticians simply don't know for sure.
Posted on March 4, 2008 10:39 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