New WSB TV / Landmark Poll has Senate Tied, Deal Ahead

WSB TV just released the results of a new Landmark Communications poll that has the Senate race in a virtual tie, with David Perdue having 47.4% of the vote and Michelle Nunn 46.6%. Amanda Swafford has 2.7%, and 313% are undecided.

In the Governor’s race, Nathan Deal leads Jason Carter 47.5% to 46.1%. Andrew Hunt has 3.5% of the vote, and 2.9% are undecided.

The poll of 1500 likely voters has a margin of error of 2.5%, which means both of the top ticket races are statistically tied. The poll’s sample is 55% female, and the African American percentage is 29.3%, ratios that Landmark President Mark Rountree has consistently stated he believes will be close to the final turnout percentages. You can view the results and crosstabs here.

In both races, around 28% of white voters favor the Democratic candidate, which would fall short of the estimated 30% needed for a Democratic win.

In down ballot races, Casey Cagle leads Connie Stokes 53%-42%. In the State School Superintendent’s race, Richard Woods holds a small lead over Valarie Wilson, 47.2%-46.5%, with 6.3% undecided. Brian Kemp is barely over the 50% mark against Doreen Carter in the Secretary of State race, 51.0 to 42.8%.

Commenting on the poll results, Rountree said,

Both Governor Deal and David Perdue have earned a slight lead in our poll going into the final weekend. Both have improved their polling numbers throughout the fall. With undecided voters removed from the equation, both are within striking distance of a majority — yet so are their opponents, who slightly trail them.


  1. therightdirection says:

    Landmark’s Trend:

    Deal: -8, -7, -4, -3, tie, +3, +2

    Perdue: -4, -7, -3, tie, tie, tie

    For this poll series, the trend is to the GOP

    • Bobloblaw says:

      While I think Landmark is correct NOW. I dont think they were right back in Aug. The trend has been for Nunn and less so Carter. Though I sense that Nunn has maximized the impact of the Outsourcing Ads and right now is stalled.

      BTW how about AG race?

  2. jh says:

    In a week, white vote has dropped 3% (66-63), black vote up 2% (30-32). Seems there’s no shortage for Nunn to mobilize. 29% black seems reasonable given demographics changes, but considering the early voting numbers, money, and interest in the race, the black vote seems poised to surprise on election day.

    • Bobloblaw says:

      Of course it is outpacing 2010. There was no competitive race in 2010 in GA.

      Nate Silver has thrown cold water on people getting excited about surges in early voting. IA and CO also have big early voting surges……because they have competitive races and didnt in 2010. Meanwhile Ohio and NV have seen a complete collapse in early voting. Why??? Cuz no competitive races. You have to make an apples to apples comparison.

      FL would be a good state to look at for early voting and comparing to 2010.

      • jh says:

        Colorado was closer in the polls in 2010 than it is today. Not trying to say anything about the total votes, just the composition.

      • Our governor’s race did not end up being competitive, but money wise and campaign wise it was competitive. In other words, we did not make it close, but it was not for a lack of trying. I think there is a distinction there. A lot of time and money was spent turning out the vote in 2010, and it worked, we just didn’t get any white votes.

  3. John Konop says:

    I am a converted Mark Rountree fan….He is king and most likely right! I was one of the first to question his polling.,… he is our local money ball guy 😉

  4. Thanks for the comments. I try to be open to questions and thoughts as they arise here on PP. I think it’s helpful for people who release public polls to do so.

    We are using the same methodology. Same weighting. If anything, the percentage of African American voters may increase, not decrease, when final votes are totaled.

    Many other polling firms have recently changed *their* weighting to show more African American voters or reduced their weighting for the number of white voters in their samples – not the other way around. We have also held firm to our gender weighting — which does affect the results since there is a significant disparity between how the genders are voting.

    Landmark has not lowered our weighting for African Americans nor female gender destination, nor changed our methodology, though we would consider it if we see real reasons to do so. At this point we still do not.

    Weighting by racial designation has an impact on polls since, of course, African Americans give 85-95% of their votes to Democrats in polls (the final results are normally actually higher for their votes Democrats than revealed in polling numbers), while white voters overwhelmingly normally support Republican candidates in Georgia.

    Same essential methodology all year long. The GOP candidates started in a hole in July due to a hotly contested primary and tough headlines, and have roller coastered up and down a bit since, but now lead by thin margins going into the final weekend (Perdue leads by less than a point, the Governor leads by 1.5%).

Comments are closed.