GA-HEG: Perdue (R) leads Nunn (D).

[Senator Cris]

http://www.examiner.com/article/new-heg-poll-georgia-deal-tied-with-carter-perdue-leads-nunn

Perdue: 47.6%
Nunn: 41.5%

[GaussLaw]

Don't know this pollster's reputation, but it seems a bit generous for Perdue.

Still, it seems like Perdue has at least a 60% chance of winning or forcing a runoff.

[Flake]

As I've said in a previous post, it really underpolls blacks and other non whites.

[Adam Griffin]

69% white? They're only off by about seven points; an accurate demographic would likely show Carter up by 4 and Perdue up by 2.

[NHLiberal]

69% white? They're only off by about seven points; an accurate demographic would likely show Carter up by 4 and Perdue up by 2.

Which sounds exactly correct

[Badger]

69% white? They're only off by about seven points; an accurate demographic would likely show Carter up by 4 and Perdue up by 2.

You're expecting the non-white share of November's electorate to be about 38%? You know GA's voting demographics far better than I, but that sounds somewhat high.

The last exit poll data I could find for GA was 2008 where CNN (fwiw) measured the white electorate share as 65% I realize the non-white population (especially Hispanics and Asians) has grown in GA during the last 6 years, but won't minority turnout drop somewhat for a mid-term election, especially among African-Americans without Obama on the ballot?

Still, 69% sounds a bit high. Maybe not impossible for a mid-term, but definitely the realistic ceiling.

[Adam Griffin]

69% white? They're only off by about seven points; an accurate demographic would likely show Carter up by 4 and Perdue up by 2.

You're expecting the non-white share of November's electorate to be about 38%? You know GA's voting demographics far better than I, but that sounds somewhat high.

The last exit poll data I could find for GA was 2008 where CNN (fwiw) measured the white electorate share as 65% I realize the non-white population (especially Hispanics and Asians) has grown in GA during the last 6 years, but won't minority turnout drop somewhat for a mid-term election, especially among African-Americans without Obama on the ballot?

Still, 69% sounds a bit high. Maybe not impossible for a mid-term, but definitely the realistic ceiling.

69% isn't possible. It hasn't been at or above 69% in a mid-term since 2006. In a state like GA, a reversion of that magnitude could only occur through genocide. We can split hairs and debate what percentage of people who are marked as “unknown” and “other” are actually white, but even that interpretation I don't think would bring it above 65%. Because we don't know for sure, though, I stick to SoS' figures.

Here's the white share of the vote as reported by the Secretary of State from 1992-2012, and an initial set of projections for 2014 to 2024 that I made last year, which I based off of averaging changes between each presidential cycle (4 years), each midterm cycle (4 years) and each presidential/mid-term (2 years). I did add a modifier of sorts to the future projections from the averages, which projected that white turnout would flat line in 2014 when compared to 2012, and going forward, would actually decrease every 2 years.

Without a Tea Party-esque wave like we saw in 2010, it's more commonplace for the white share of the electorate to tick up 1 to 1.5 points in a mid-term when compared to the previous presidential (it actually decreased between 1996 and 1998). I'm a bit more bearish than I was when I made this graph last year, though, and as such, I'm assuming that max white turnout will be somewhere between 62 and 62.5%.

With a 2010-style wave that takes into account demographic shifts over the past four years, it could be 64% (which is what a lot of the campaigns are assuming when conducting their turnout calculations). While my 2014-2024 projections on the graph may be slightly optimistic, the past trends show us that there hasn't been an instance where the mid-term white share spiked to the same total as the previous midterm while also seeing a substantial reduction in white voters in the presidential cycle between them.

[Badger]

69% white? They're only off by about seven points; an accurate demographic would likely show Carter up by 4 and Perdue up by 2.

You're expecting the non-white share of November's electorate to be about 38%? You know GA's voting demographics far better than I, but that sounds somewhat high.

The last exit poll data I could find for GA was 2008 where CNN (fwiw) measured the white electorate share as 65% I realize the non-white population (especially Hispanics and Asians) has grown in GA during the last 6 years, but won't minority turnout drop somewhat for a mid-term election, especially among African-Americans without Obama on the ballot?

Still, 69% sounds a bit high. Maybe not impossible for a mid-term, but definitely the realistic ceiling.

69% isn't possible. It hasn't been at or above 69% in a mid-term since 2006. In a state like GA, a reversion of that magnitude could only occur through genocide. We can split hairs and debate what percentage of people who are marked as “unknown” and “other” are actually white, but even that interpretation I don't think would bring it above 65%. Because we don't know for sure, though, I stick to SoS' figures.

Here's the white share of the vote as reported by the Secretary of State from 1992-2012, and an initial set of projections for 2014 to 2024 that I made last year, which I based off of averaging changes between each presidential cycle (4 years), each midterm cycle (4 years) and each presidential/mid-term (2 years). I did add a modifier of sorts to the future projections from the averages, which projected that white turnout would flat line in 2014 when compared to 2012, and going forward, would actually decrease every 2 years.

Without a Tea Party-esque wave like we saw in 2010, it's more commonplace for the white share of the electorate to tick up 1 to 1.5 points in a mid-term when compared to the previous presidential (it actually decreased between 1996 and 1998). I'm a bit more bearish than I was when I made this graph last year, though, and as such, I'm assuming that max white turnout will be somewhere between 62 and 62.5%.

With a 2010-style wave that takes into account demographic shifts over the past four years, it could be 64% (which is what a lot of the campaigns are assuming when conducting their turnout calculations). While my 2014-2024 projections on the graph may be slightly optimistic, the past trends show us that there hasn't been an instance where the mid-term white share spiked to the same total as the previous midterm while also seeing a substantial reduction in white voters in the presidential cycle between them.

Thanks for the stats, Adam. I didn't think to check the SoS as Ohio's doesn't to my knowledge maintain figures for estimated racial breakdowns of the electorate.

You've more than convinced me the white share won't hit 69%. I think your estimation for the mid-term increase in white share is a little "flat" compared to previous years, but I agree the ceiling should be somewhere in the 64-65% range. The most likely is probably somewhere in the low to mid 60's.