11Alive Survey USA Poll Says Races Have Tightened

11Alive and Survey USA released a new poll this morning showing the races at the top of the ticket have tightened.  In a poll conducted yesterday, the results are:

Nathan Deal 45%
Jason Carter 44%
Andrew Hunt 4%

David Perdue 47%
Michelle Nunn 44%
Amanda Swafford 5%

In other races, the GOP fares better:

Casey Cagle over Connie Stokes 51% to 40%
Brian Kemp over Doreen Carter 52% to 39%
Sam Olens over Greg Hecht 49% to 41%
Richard Woods over Valarie Wilson 47% to 43%

Survey USA claims they used identical polling methods between yesterday’s poll and one a few weeks ago that had much larger margins.  And yet, the makeup of the respondents is significantly different.  Specifically, yesterday’s poll included 29% African American respondents, versus 26% August 14-17.  Hispanic participation dropped from 7% to 6%.

Mike Hassinger would like me to remind you that polls are for strippers and cross country skiers.  Trite political discussion requires me to remind you that the only poll that matters is on election day.  But still, argue over this one like your life depends on it below.


  1. John Konop says:

    Casey Cagle look like he will be a very strong statewide candidate for Governor in the future…..Brain Kemp looks strong as well for other statewide offices….

      • John Konop says:

        In fairness you have been consistent about predicting a runoff. But Mark is right about the trend….the issue is just how far…..Mark did show guts and integrity for being the first…..When Mark did this polls all other polls showed major lead for GOP. You have to give him a little “that a boy”…

      • And Bob, they still haven’t used a data weighting that is comparable to ours for gender…they said it was 51% when it was 55% in 2010 (the last governor election)…all while reporting that women support Nunn by double digits. Landmark at the time had males 47-38 for Deal, and this one by SurveyUSA has males 50-38 for Deal (quite similar); Landmark at the time had females 34-50 for Carter, while the SurveyUSA poll also had 50% of females for Carter to 40% for Deal (slightly higher, three weeks later than we polled.

        See for yourself the cross tabs:


        Actually, the numbers are strikingly similar, but still weighted differently. Landmark weighted more for women to vote (more in line with the 2010 election), and we provided what I believe to be a broader option for self-describing one’s race:

        SurveyUSA didn’t allow a self-described racial option for asian and “other”, though I am unsure as to how this might impact the results (do non-black minorities not complete and therefore are not counted? Don’t know…)

        You asked for reasons for the differences. Now you see them.

  2. While I suspect they might have put their thumb on the scale a little, it is actually possible for their methodology to have pretty wild swings in the makeup of the electorate as they use an intensity scale to compose it.

    Really that just shows that their methodology has problems especially when polling far out from an election.

  3. Jon Lester says:

    I come here most days, so I’m probably better-informed than most Georgia voters, but I honestly forgot that Cagle, Kemp and Olens are up for re-election, and I don’t know anything about their respective opponents, except that they, too, appear to have been hung out to dry by the state Democratic party. I would have expected a tighter race for SoS, at least.

  4. Monica says:

    Not that it matters, but I believe the post above has it wrong placing Deal ahead by that one point. The linked article to 11Alive shows that it is Carter who is ahead. “Forty-five percent of registered likely voters plan to vote for Jason Carter, 44% for Nathan Deal. The margin of error is +/- 4.2%, so they are statistically tied.”

  5. Here is Todd Rehm’s comment on the fact that the Channel 11 pollster, SurveyUSA, suddenly changed their demographic weighting to be more in line with the Landmark weighting (as well as Todd’s own poll weighting at GAPundit)

    (Todd’s comment below)
    “11Alive has a new poll out they claim to be “identical” to three weeks ago, but now they’ve boosted their estimate of African-American turnout from 26% to 29%, more than a ten point increase. It also brings their prediction of turnout in line with those made by Landmark’s poll and GaPundit’s poll. This change is significant and causes part of the change in the head-to-head ballot numbers for Governor and Senate.”

    –Todd is correct that they have changed their weighting. This is exactly what I have been saying on behalf of our polls at Landmark, as well.

    This is important because of the disingenuous reaction to our last survey by some politicos who didn’t know better.

    Or did.

    • Bobloblaw says:

      So Mark where is the Nunn 6 point lead here? My criticism of your poll wasn’t the weighting, it was that the poll was likely the 1/20th polls that is statistically wrong. Still no poll has matched your poll from THREE weeks ago. It is really amazing your claiming vindication when the spread is a whopping 9 points between this poll and yours for the Senate race.

      With Swafford/Perdue we are at 52%.

  6. Bobloblaw says:

    It will be interesting to see if the indications of a little GOP wave begin to impact these polls in GA. The GOP has opened up a small but stat sig lead in the national generic ballot at RCP. Usually individual races trail the generic ballot by a few weeks. I predict that nearly all close Senate races . will break the same way as they do all in most years (except 2010 thanks to bad candidates). Cotton has moved into the lead in AR, The Fat Woman is close to singing in KY and Sullivan appears to have opened a lead in AK. I predict that NC will be one of the hardest races for the GOP. That even as others open up for the GOP, NC wont or it wont until later. Orman will fade in KS, plus who knows who we would caucus with anyway.

    • Harry says:

      Furthermore, Republicans are broken glass folks who will turn out on election day. The Democrats will not be motivated to get up and vote for their white candidates. Just the facts.

      • Bobloblaw says:

        Im glad you brought that up because I was going to post a stand alone post on something like that, but instead Ill just reply.

        Im not sure it is right this year, unlike 1994 and 2010.

        Look at the poll numbers. What is consistent among all the races is the Dems numbers(low to mid 40s). What is not consistent is the top of the ticket and the bottom of the ticket on the GOP side. Perdue and Deal are struggling, while Olens, Kemp and Cagle are hitting 50%. (contentious GOP GOV primary for 2018?) Some good news is that the undecideds in the Perdue and Deal races are likely GOP voters (people voting Olens, Kemp, Cagle). So there is more upside for Deal and Perdue than Nunn and Carter. Problem is Deal and Perdue had contentious primaries. This makes me wonder. Are there supporters for Kingston/Gingery/Broun/Handle and Pennington who aren’t yet supporting Perdue and Deal? But they are onboard for Cagle, Kemp and Olens? TP voters are very cantankerous and might well leave Deal and Perdue blank to “teach a lesson to the GOP”. I don’t know. But the above numbers tell me Perdue and Deal are likely suffering from rough primaries. But likely the undecideds are GOP voters. How disgruntled they are is the question.

        Go read Freerepublic.com some time to see just how marginalized the far right feels by failure to defeat McConnell, Cochran and Alexander. That site is full of hopping mad TPers who are rooting for Grimes over McConnell to teach the GOP a lesson, whatever that may be.

        • caroline says:

          So far this year is reminding me of 1998 more than anything. Obama is going to be gone in two years and people like Perdue are asking to be elected for 6 years. What does he stand for other than not liking Obama? And the GOP just seems to be so certain that talking about Obama like they did about Bill Clinton back in 1998 is going to carry the day and they aren’t going to have to actually give people something to VOTE FOR.

  7. Mrs. Adam Kornstein says:

    A pissing match over a tight poll, quelle surprise here at PP.

    This is where the ephemera of campaign season kicks in hard.

  8. caroline says:

    The Gov and senate races have really been interesting to watch. It seems Perdue has completely given up on trying to win any moderate votes and gone completely far right with his advertising. One of his ads was so over the top my husband and I burst out laughing watching it.

    The Deal/Carter numbers are more surprising to me. I have to say though Deal’s advertisement saying that things are great makes me wonder if he thinks everybody in the state is stupid. Last time I checked Georgia was second highest in unemployment in the nation.

    • Harry says:

      Perdue has to go right in order to win back the GOP base. He isn’t planning to get many “moderates” of which there are few in Georgia anyway, so there is more net benefit from an appeal to the base.

      • caroline says:

        Well, I guess we’ll see if it works or not. That is pretty much what I gleaned from his advertisements but I’m not sure that there aren’t many moderates in GA though.

        • Harry says:

          There really aren’t that many “moderates”. Right-thinking independents and libertarians yes there are many; but the few claiming to be “moderates” are actually Democrats.

            • Harry says:

              Very few. There are far more right-thinking independents and libertarians that need motivation, and running anti-Obama ads is one sure-fire way to do it.

              • Bobloblaw says:

                I think the numbers actually may show youre correct. Someone is supporting Cagle, Kemp and Olens but not Perdue and Deal. My guess is it is TP folks angry with the primaries. I may be wrong. Deal and Perdue WILL win. The question is runoff or not. Right now looks like runoff. That favors the GOP.

Comments are closed.