Last week confirmed my belief that the polls are providing more political interest than several of the candidates in the major races in Georgia. Political junkies were barraged by the release of one poll after another in the races for governor and U.S. Senate.
With my time divided between Florida and Georgia, I now realize the contrast between Florida’s gubernatorial contest in which every moment of TV advertising is filled with ads for or against Gov. Rick Scott or former Gov. Charlie Crist. Florida is red hot with action from its candidates. Georgia, on the other hand, is left to entertainment by… polls.
When I suggested long ago that the gubernatorial race could go to a runoff, at least one candidate and his team laughed at the possibility. Now, with endless polls showing Gov. Nathan Deal and Jason Carter virtually tied in the low to mid forty-percent range, I would guess the laughter has subsided. Deal’s campaign ads are matched only by Carter’s for their unimaginative and clearly uninspiring message.
While Michelle Nunn and David Perdue and their backers are supplying enough fireworks to be “Florida worthy,” the two gubernatorial candidates seem to be imitating the Atlanta Falcons who, during most years in team history, waited until they were down 21 points and had six minutes to play, to start really hustling (or in modern times, going to the no-huddle offense). September is slipping away and the messages of economic growth from Deal and education woes from Carter are as old and tired as some of those past Falcons playbooks.
So the political pundits are left digging for little nuggets to fill space and an avalanche of polls to pick apart. So, let’s start picking.
First the easy one. Deal has remained at around 44% in the RealClearPolitics average of all polls and Carter is around 42%. I’ve seen it said by some pundits that if an incumbent governor is below 50% he or she is considered to be in trouble. That old rule of thumb actually died some 10-plus years ago. These days the measuring stick has moved down to the forty-five percent or so mark. But even at that level, the governor would be considered to be in trouble. And if Carter holds Deal to that level going into October then Deal will move from a status of being in “serious condition” to being nearly in political “extremis.”
The variables that will decide if we head into a runoff are the following:
* Will Andrew Hunt, the Libertarian candidate, be able to hold at least three and a half percent of the vote? That’s the level where tepidly run races coupled with strong turnout by African-American voters can move a Republican incumbent into a runoff position.
* Will Nathan Deal’s campaign strike with their secret weapon? Governor Deal’s wife is a warm and down to earth First Lady who could help lure older female voters, who have watched one too many TV news stories about ethics and the like, back to the Deal camp.
* Will the Carter campaign take the ethics questions to those under fifty years of age who don’t read newspapers and don’t watch local TV news at the same avid level of senior voters? To do so requires avoiding their desire to put their candidate and typical Democrat education rubbish on air. For all of his smiles that Jimmy Carter flashed, he knew when to get tough in a campaign. For his grandson, time is running out.
* Will Carter’s youth work for or against him in the debates the two candidates will have? And does anyone still watch them?
Now to the more complicated race. As of early Friday night, our InsiderAdvantage poll seemed “the outlier” of the polls released last week. While we were right in the sweet spot of other polls as to the Deal-Carter race, we appeared off in the Senate contest. That was until later that evening a Landmark poll showed both Nunn and Carter leading their respective Republican opponents. No other poll released last week, including the AJC’s survey, which I would consider to be the best “average” of all of the surveys released recently, had Perdue and Deal with leads.
While our survey which showed Perdue at 50% compared to Nunn’s 40% showed the widest lead we have seen in the race for Perdue, it seemed about on target—not only with regard to the Deal-Carter contest, also with regard to President Obama’s approval rating in Georgia which it pegged at 41%. Ironically our survey came closer to numbers produced by CBS and the New York Times in the Nunn-Perdue contest. Who would have guessed that?
The AJC and SurveyUSA polls seemed reasonable to me, and even the Landmark poll could be accurate. So am I saying none of us know what we are doing? Sort of!
Polling Georgia is tricky business because of our unique mix of rural versus city, natives (a shrinking number) versus transplants, young versus old, and the strong number of African-American voters and a growing number of other “non-white” residents.
Our firm has decided to weight African-American voters at around 33% of the expected vote this November. This is an adjustment we made based on several factors. First, the African American turnout in 2010—the year of the Tea Party—was nearly 30%. When determining the weight for one demographic we must consider the weight to be applied to other groups. We have reason to believe that the turnout among white voters will be strong this year, but no more than in 2010. With several of the counties with the largest African-American populations moving to a “Sunday” vote and a major registration effort underway, we believe that the historic levels of black voter turnout seen in 2008 and 2012, as a percentage of the overall vote, might actually be eclipsed this year. No pollster can remain stuck on their models, so as data or events change, so might our weighting. But it is very hard to imagine that we will be weighting African-American voters under 30% at any future point in this election year.
Then there is the issue of data collection. The landline telephone is dying quickly, but not so quickly that it can’t provide a polling sample. Mobile phones, smart phones, and desk top computers are quickly replacing landlines as the best source for gathering a strong sample of voters. But everyone is all over the board as to the very best way to approach polling.
The combination of IVR (automated) phone calls coupled with online responses actually proved to be our best means of collecting accurate data during the GOP Senate primary and runoff. The polls in which we utilized that combination (we dubbed them to be a “Super Poll”—creative, huh?) actually pretty much nailed the two races. Our IVR standalone polls were more likely to be off a bit, but within their margins of error. Polling that combines IVR with mobile devices is our current system of gathering responses. We likely will look at that model and alternate between that approach and the “Super Poll” system.
There is absolutely no way to tell which of the many pollsters in the Georgia contest are on target. My best suggestion to the hardcore political junkie is to constantly look at the RealClearPolitics average in both the gubernatorial race and the battle between Nunn and Perdue. At the moment, that average would show Deal and Perdue leading their respective races. But those averages would also suggest that the days of Republicans rolling into office with just a mediocre effort are over. I mean really over. And the days that the polls are more active than some candidates have arrived.
