WXIA Poll Shows Tight Race for Senate and Governor

A new poll released today by 11 Alive shows a tightening election, both at the top of the ticket and down ballot.

In the race for Governor, Democrat Jason Carter leads Republican incumbent Nathan Deal by a point, 45%-44%, with Libertarian Andrew Hunt taking 4% of the vote. In the Senate race, it’s also a one point margin, with Republican David Perdue leading Democrat Michelle Nunn 46% to 45% and Libertarian Amanda Swafford picking up 4%.

The biggest change from a similar poll two weeks ago occurs in down-ballot races. In the race for Secretary of State, a seventeen point lead by incumbent Brian Kemp over Democratic challenger Doreen Carter has dropped to two points. A twelve point lead by Richard Woods in the State School superintendent race has also dropped to two points against challenger Valarie Wilson.

The poll’s margin of error is 4.3%, so all of the above results are within the margin of error.

But looking at the crosstabs, the sample appears to favor Republicans, with a 50-50 split between men and women voters, and only a 27% black vote. Women normally vote in higher percentages than men, and the 27% black vote is on the low side of many other estimates.

The percentages also vary from the poll taken two weeks ago. Here are the crosstabs from that poll. Black voting strength is down by two points, and female voting strength is down by a point.

With the usual disclaimers, what should be made of this poll? Is bobloblaw right? Tell us.


  1. There goes Bobloblaw’s thesis. I still think Cagle at 47 (+4) is in infinitely better shape than Deal at 44 and -1.

    Based on this poll, I’d say odds are something like:
    Governor: Carter win 35%, runoff 60%, Deal win 5%
    Lt Gov: Cagle win 70%, runoff 30%

    • David C says:

      Like I said before: Republicans in other statewide races could only use if they end up on the news with a big scandal that would make voters actually pay attention to them. Cagle isn’t, Kemp is.

    • Bobloblaw says:

      There is not a 60% chance that Carter will win in a runoff. If it goes to runoff in both races, it favors the GOP.

      • David C says:

        Seems pretty clear Chris is saying that there’s a 35% chance Carter wins on election day, 5% Deal wins on election day, 60% chance the election goes to a runoff, not that there’s a 60% chance Carter wins it.

        • Bobloblaw says:

          A runoff is far more likely than 60%. Fact is Carter isn’t polling much better than the overall Dem margin in the past 10 years.

          • jh says:

            Sure, up 1 or a few more in SurveyUSA/Landmark isn’t doing better than 2010, when they were down 10 at the same point in 2010.

            • +1 jh. You are right.

              When you revise the various national polling firms’ demographic weighting problems to reflect a more likely voter turnout of 29% black and 55% female (instead of 25% and 51% respectively, which most national pollsters seem to be using), those polls then reflect Carter leading or tied with Gov. Deal.

              At this stage in 2010, Deal led Barnes in every poll (Rasmussen showed +6, Insider +8). Landmark had the race at this stage by a margin of +8 for Deal. Deal remained ahead in all polls by all firms throughout the rest of the 2010 election by leads that were outside the margin of error.

              But in 2014 that lead no longer exists. When proper demographic weighting is applied to nearly any poll by any firm, the race has Carter leading or tied. By no measure is 2014 similar to 2010. There are many more minority voters registered today than in ’10, and among white voters there is a strong current of anti-incumbency. Anti-incumbency is a political problem facing Governors all over the USA of both parties.

              Bobloblaw, you can’t have it both ways. You can’t simultaneously say that polls in 2014 are similar to polls in 2010 –while also saying that there will be a runoff. The polls are significantly different at this stage, and have been for some time. Female voters are voting for Democrats by double-digit margins at this stage (a 26% net differential between men and women). You seem to be trying to cover your bases, and it’s pretty transparent.

              • Bobloblaw says:

                Im am saying that the Dems arent polling any better than in the past. It is the GOP that is polling worse. Undecideds are most likely GOP leaning voters.

                • Will Durant says:

                  What you are consistent in saying is that your glasses are not just rose tinted, the are ruby red. Republicans are better served if they take a realistic look at the numbers and act accordingly. Part of that strategy should be to get SoS Kemp to shut the Hell up, for example. He’s firing up the wrong base as far as the Republicans should be concerned.

                  • Bobloblaw says:

                    Nunn leads in only one poll, Carter leads in about 1/4 of the polls. Most people arent paying attention to Kemp. Minority turnout will not be historical in 2014.

  2. jh says:

    Females are 50% in this poll when they are usually 55%.

    I think it’s great InsiderAdvantage publishes the ridiculous polls giving Perdue ahead, keeps Republican money away from the race. Not that the terrorist ads are working.

    • FranInAtlanta says:

      National Republican money running ads that tie Nunn to Obama help Nunn with a substantial number of voters. I suspect they help her more than they hurt her.

      • jh says:

        The black vote is going to be the black vote. The white vote is the biggest variable, and it looks like it’s firming up for Carter/Nunn.

  3. Very good catch, Jon Richards. The weighting once again is off based on historical records.

    56% of Georgia’s voters are female, yet this poll only has the female vote at 51%. 55% of the voters in the 2010 gubernatorial election were female — and the state is even “more” female in its composition today. Women voters are currently going for Democratic candidates by double-digits in 2014, though this could change.

    And in reviewing turnout projections by race, this one shows 27%, while the more historically accurate number would be about 29%-30%.

    This racial weighting adjustment is important for accuracy because the 2% difference would essentially all go to the Democratic statewide candidates.

    For accuracy: these two changes would change things approximately as follows: racial adjustment adds +2 to Carter/Nunn, and the gender adjustment would add +1 or 2 for them as well.

    With these simple demographics weighting adjustments for accuracy, the SurveyUSA numbers now show exactly what we are showing at Landmark. But in the end it’s probably a runoff for both races.

    This exercise is important because candidates and a political party might have based their communication and operations off of these kind of inaccurate polling numbers.

    I think we’d be seeing different campaigns, messaging and advertising today if these numbers had been honestly acknowledged earlier in the year.

      • The ultimate problem for Deal is like what benefit do you get from Deal? Even if you want the Republicans to be in control of state government, they effectively still will be with 65% of the legislature. But there’s no way to rein them in. At least with Carter you get a check.

        Let’s step out of politics and into the reality world for a second. On an issue like Medicaid expansion, it is OBVIOUS to all viewers who aren’t hopelessly partisan that Obamacare is here to stay, that Medicaid expansion makes sense as at least the first step, and that a majority of Georgia supports expanding Medicaid.

        Yet Deal has to pretend that he can never do it because of partisan politics. Meanwhile, Carter could probably take his lumps, forge some sort of “compromise” and get on with reality.

        All a bonus – no downside to the state/average voter.

        Now of course, this analysis assumes that people like Deal – there’s really no upside if you like Deal. If you think he’s a crook, well then there’s downside to voting for him again, and a LOT of upside to giving Carter a chance.

        • John Konop says:

          I think Gov. Deal has done some very positive reforms for the state ie vocational education, prison reform, promoting the state…….I agree the ACA Medicare issue is a problem…..but the solution is not so simple…unless we bend cost curve you will just add more people without doctors to serve them……

          • Medicaid (especially for the type of mostly healthy working poor people who would be covered by expansion) is not where the cost curve problems are.

            And it’s not like we don’t have data from 20 other states who did expand Medicaid to compare to, I don’t think a single one can show that they’re worse off for doing it yet.

            • John Konop says:


              In all due respect the reality is non of it is sustainable unless cost curve bends….. I have made numerous suggestions in the past from dial doc for non emergency care, phones in waiting rooms for non emergency care, increase access to birth control for low income people, mandatory living wills, use VA pricing for drugs…….


              …..The Unsustainable Cost of Health Care – Social Security Advisory …
              driving up the cost of Medicare and Medicaid to unprecedented levels. … If these
              costs continue to rise as rapidly into the future, the standards ….. spending is
              concentrated on end-of-life care and that much of the problem with the health….


            • Bobloblaw says:

              Just wait. If Georgians are upset because we don’t have enough welfare, then the state is finished as a dynamic state economically.

          • Bobloblaw says:

            “””..but the solution is not so simple…unless we bend cost curve you will just add more people without doctors to serve them……”””

            That is what I’ve been saying. Liberals think access to insurance is access to care. Without increased providers, expanding Medicaid does nothing to expand access.

        • David C says:

          That point: That a Gov. Carter could be a check on a guaranteed highly Republican legislature, is a good one: It’s how Republicans have long been able to win Governorships in otherwise deep blue Northeastern states like Massachusetts, Rhode Island, New Jersey, and Connecticut, and Democrats in places like Wyoming, Oklahoma, and Kansas.

        • Will Durant says:

          Don’t forget the legislature voted in an inoculation just in case a Democrat was elected this year. They put the control of Medicaid expansion or healthcare exchanges in their hands, not the Governor’s.

        • Bobloblaw says:

          A majority of GA will support Medicaid expansion until they realize they will have to pay for it at some point. Also expanding medicaid without expanding providers will simply create massive shortages in the healtcare system.

    • Bobloblaw says:

      As usual the pollsters are right everywhere but GA right? GA is sooooo different than every other state right? If you start monkeying with the crosstabs to the extent you do, youre doctoring the poll. You sound amazing similar to the Unadjusted Polling website in 2012, that took all the polls and adjusted them to show a Romney victory.

    • Bobloblaw says:

      You’re ignorng the fact that the poll also shows only 63% white support for the GOP. It will be closer to 70%. So that cancels out your point. I don’t worry about demographics because there is enough offsetting data to maje the overall result correct.

  4. northside101 says:

    You have to figure Nunn is certainly guaranteed from the start (a floor, not a ceiling) the 39 percent who backed Michael Thurmond in his suicide run against Senator Johnny Isakson in 2010. As for Carter, he ought to be able to match the 43 percent Barnes got against Deal in 2010.

    The Democratic strategy for victory might simply be described as “30/30”—get turnout to where blacks are 30 percent of the statewide total turnout (as was the case in 2012) and get 30 percent (perhaps a little more) of the white vote, and you have a majority, even if slim. Offhand, I don’t think Obama got even as much as 25 percent of the white vote in Georgia in either 2008 or 2012—23 percent seems to stick in my head—but even when he surprisingly lost to Perdue (Sonny, not David) in 2002, Barnes managed about a third of the white vote. Problem was for him, that was insufficient to win as blacks back then accounted for only 23 percent of the state’s total turnout.

      • Exit polls are notoriously buggy (the age #s don’t make any sense in the ’08 one) but the race #’s seem pretty good on this one.

        Of course it’s shameful that they cut Georgia out of exit polling, as it’s the most likely red state to flip blue (behind North Carolina which depending on your definition is or isn’t still red).

        It’s great that they’re exit polling Texas – a state the Republicans have no chance of losing for 15 years.

    • David C says:

      Given that Romney supporters were trying to reweight by (ultimately specious) thoughts on party ID, while much discussion here is on the composition of the electorate in Georgia’s relatively racially polarized voting environment, the comparison between this and the “unskewed polls” guy is pretty distant. Most pollsters reweight their samples to make them look more like their vision of the electorate, especially as response rates go down in a world where fewer and fewer people have land lines and answer polls. What Rountree and Chris are both saying is that the way they think the electorate will look is different than some of the other pollsters. Something like party ID can be flexible (especially between people who say they belong to the party and say they’re an independent while voting a solid ticket anyway) while obviously things like race and gender aren’t. So saying Georgia’s going to have a 30% African American electorate, or a 55% female electorate and you need to weight accordingly are reasonable judgments that you can look back on recent precedents and Georgian demographic information. The guy who said he needed to reweight all the polls to put more Republicans in them because he was convinced 2012 was going to be 2010 all over again, not so much.

      • Yeah that’s exactly right. We’re lucky in Georgia, it is a relatively easy state to poll you just have to get the composition of the electorate correct, which many seem to have a problem with.

        The AJC poll is a good example – by all accounts their field work was quality, they just goofed when they created their likely voter screen that had only 24% black respondents make it through and equal numbers of men and women.

        Now some others like Insider Advantage don’t make any sense at all – especially their polls that have shown 65+ voters being more Democratic than the electorate as a whole. This is demographically impossible.

      • Bobloblaw says:

        But they are very convenient in their analysis. They look at the demographic breakdowns but not the results within each breakdown. For instance whites giving Perdue 63% support, would be a historical low. Thus the underweighting of blacks and women is offset by the lower percent white going to Perdue and Deal.

        Mark and Chris act like GA is the only state with a sizeable minority population. If you believe them, we are going to see stunning historical minority turnout. Beyond anything predicted by anyone, heck lets just call it what Chris wants to call it. Majority turnout. Exceeding even 2008 percentages.

        You can tell Chris’s shrill bias. Even if you buy Landmark’s poll, Nunn’s lead went from 6 to 3 in 3 short weeks. Looks like Perdue is doing something right. Youd think that might worry Chris, the only poll Nunn has lead in shows her lead cut in half. Id worry.

        • Since I’m not an idiot I am also able to look at my tv and notice that during the 6 point lead, Nunn was owning the advertising space to herself, more recently Perdue has been advertising.

          More simple explanations to Bobbloblaw objections!

        • No, Bob, again you are wrong and pressing your hysteria button.

          It’s not a question of historical turnout for the black vote, just regular-ole’-turnout-plus-four-years that is projected.

          –In 2008, the year that Obama energized the black vote across the country, 30% of the electorate was black.

          –In 2010, 28.5% of the electorate was black. There is some issue with the ‘stated’ black % number by the SOS, which was 28.25% because that didn’t include ‘unknown’ voters, which are also more likely to be minority.

          –In 2012, the black vote again was right at 30% (barely under). However, again with the unknown race vote added, it was over 30%.

          There is no historical turnout projected, other than the fact that the black vote in Georgia continues to climb and will almost certainly be higher than in 2010 (28.5% black), which was the low point for Democrats in Georgia.

          The percentage of the black vote in Georgia has increased steadily since 1996 when 19% of voters were black. Essentially the electorate gets about 1% net more black with each passing election. Plus non-black minorities, which likewise give significant vote majorities to Democratic candidates, are also gaining in turnout strength.

          The Republican Party **must** make inroads with minority voters to retain its statewide majority position. The GOP do not need to emulate Democrats on issues to do this: they need to recruit and most importantly *nominate* black candidates and other minority candidates.

Comments are closed.