New Poll: Perdue, Kingston Lead on First Day of Early Voting

April 28, 2014 17:13 pm

by Jon Richards · 16 comments

David Perdue leads the Georgia Senate race with 26% of potential voters, while Jack Kingston has 20%, according to a new Survey USA poll released by WMAZ in Macon. Karen Handel polls at 15% and Paul Broun has 13%, making the two statistically tied, as the poll has a 4.5% margin of error.

The poll also predicts Governor Nathan Deal will defeat challengers David Pennington and John Barge without a runoff.

Complete results and crosstabs here.

Early voting for the May 20th primary began today. Discuss.

{ 16 comments… read them below or add one }

George Chidi April 28, 2014 at 9:19 pm

In the Democratic primary, how in the hell is Steen Miles polling at 13 percent while Branko Radulovic gets 5 percent? If those numbers pan out, I’m firing myself as a judge of the electorate.

Mike Hassinger April 28, 2014 at 10:39 pm

George, Dr. Rad has not been on TV for as long as Steen Miles, and I seem to recall she also ran for something before -State House? DeKalb CEO? So there’s your answer.

Will Durant April 28, 2014 at 11:04 pm

C’mon George, you haven’t seen all of the comments in this forum about the dangers of electing politicians with a higher education? Don’t forget Mencken when you are judging the electorate.
And I agree with Mike that you should never discount name recognition either of course. I would be willing to bet that Dwayne Elizondo Mountain Dew Herbert Camacho would have polled higher than Branko Radulovic had he been one of the choices.

NorthGAGOP April 28, 2014 at 10:12 pm

Perdue spends $2,000,000 on TV and is sinking, Kingston $1,500,000 on TV and hasn’t moved, Handel $.20 on TV and rising. Gingrey and Grayson tied. This is going to be a barn burner.

TheEiger April 29, 2014 at 10:57 am

So it seems that Survey USA still has not figured out how to do a poll in Georgia. They have once again under sampled Atlanta by 10 points. Which means Handle and Kingston are more than likely in a dead heat at 20%. Perdue has spent $2 million on the crying baby ads only to drop 3 points. Somebody should be worried.

analogkid April 29, 2014 at 6:39 pm

I don’t purport to be a polling wizard, but here are a few thoughts:

First, you’re correct that nearly 33% of the population of Georgia lives in Cobb, DeKalb, Fulton, and Gwinnett, but the poll is suggesting that 25% of likely GOP voters reside in those four counties. For the Dem primary, they estimated that 40% of likely voters reside in those same four counties. That sounds about right to me.

That said, they do acknowledge that Handel has “momentum,” so you may end up being right about the result even if Survey USA’s methodology is not actually flawed.

TheEiger April 29, 2014 at 9:06 pm

You are leaving a few counties out of what is considered “Atlanta”. Henry, Paulding, Cherokee, Rockdale, Walton.

analogkid April 29, 2014 at 9:35 pm

That may be your definition, but it isn’t Survey USA’s:
http://www.surveyusa.com/SUSA_Regional_Definitions_As_Of_081029.htm
(Click the GA tab at the bottom)

TheEiger April 29, 2014 at 9:47 pm

They are still under sampling Cobb, Fulton, Dekalb and Gwinnett 5 points and over sampling south GA.

analogkid April 29, 2014 at 11:19 pm

OK, so I did a little digging on census.gov and just shy of 27% of the entire white population of Georgia lives in Cobb, DeKalb, Fulton, and Gwinnett. If every Georgia citizen voted, and every white person in those four counties pulled a Republican ballot, then the white voters would obviously represent 27% of the total electorate.

But not everyone can (or will) vote, right? It’s possible that folks that live in certain geographic areas or that subscribe to certain party affiliations are less likely to vote in primaries than others (I honestly don’t know), but if you assume that’s fairly consistent across the state and party affiliation, then the absolute high water mark for Republican primary voters in those counties couldn’t possibly exceed 27% of the entire state.

And that’s before you consider the fact that white people in urban areas are way, way more likely to be Dems than anywhere else in the state. Even if you assume that some blacks, hispanics, and asians will pull an R ballot, it’s nowhere near offset by the number of whites that will pull a D ballot.

Someone check my math, but in light of the above, I think Survey USA’s 25% figure of R primary voters for the Atlanta region is, if anything, too high.

TheEiger April 30, 2014 at 7:26 am

Okay, I’ll hold you hand on this one and walk you through it. You are looking at census data and not voter data. I would encourage you to go to the Georgia SOS website and look at historical voting data. In the 2010 Governor’s race those counties made up 27% of republican primary voters. In the 2012 presidential primary those counties made up 29%.

See the trend upward? As I said they have under sampled those counties. I would encourage you to look at the correct data in the future.

analogkid April 30, 2014 at 7:30 am

You beat me to it. See below.

analogkid April 30, 2014 at 7:30 am

A better use of my time last night would have been to just look at actual election data…

Using the exact same regions as Survey USA:

2010 Primary:
27% Atlanta, 45% Northwest, 28% South & East

2012 Primary:
29% Atlanta, 44% Northwest, 28% South & East

Compare that to Survey USA’s breakdown for this poll:
25% Atlanta, 42% Northwest, 33% South & East

So, I’m inclined to agree that they slightly underestimated Atlanta and the Northwest, which would result in an overestimate for the South & East.

JayJacket April 29, 2014 at 11:27 am

@TheEiger: I was also wondering about this. Their sample (25% ATL, 42% NW, 33% South/East) would seem to give Perdue and Kingston artificially high numbers while hurting the Atlanta-centric candidates (i.e. Gingrey and Handel). Not sure about the impact on Broun.

Demonbeck April 30, 2014 at 8:42 am

All of this is very interesting, except for the fact that those “Atlanta-centric candidates” are losing to Purdue and Kingston in the Atlanta market as well.

TheEiger April 30, 2014 at 10:08 am

If Survey USA can’t figure out how to get the sample right you should pretty much dismiss everything else as well. The link below shows the polls taken during the 2010 Governors race. Both Survey USA and Magellan did a poll during the same period of time. Survey USA had Oxendine ahead by 9. Magellan (they actually predicted the real outcome) had Handel winning by 14. What I’m saying is Survey USA is worthless in their polling in Georgia.

http://www.realclearpolitics.com/epolls/2010/governor/ga/georgia_governor_republican_primary-1332.html