Georgia senate seats runoff(s) megathread

[Adam Griffin]

With regard to racial composition of the runoff electorate thus far, remember that early in-person voting begins today and the GOP electorate will start pouring into the ballot boxes. For November's election, VBM racial breakdowns had whites in the low-50s before early in-person voting began (just as they are now); by the time it ended, whites were back around 58-59%. In a normal election, the electorate would begin getting substantially whiter with each passing day. That effect may be muted somewhat for a runoff, since this election has naturally lower awareness among rank-and-file voters; a greater percentage of black voters may also vote in-person this time when compared to November.

Also remember that among the mail ballots requested, 631,332 were automatically sent due to requesting a mail ballot in the primary and/or general by those 65+, disabled, veteran or overseas (the 65+ segment is the vast, vast majority). That's a majority of all mail ballots requested for the runoff thus far, so don't read too much into the total number of mail ballots requested (especially because of the age breakdowns among these voters). Likewise, with only 21% of requested ballots being returned thus far with early in-person voting starting today, these VBM request totals are pretty worthless as far as #analysis goes.  

[Adam Griffin]

Higher in-person turnout relative to November is only indicative of the reality that fewer likely voters will be voting by mail in this election (again, do not read heavily into the overall requested mail ballot figures for enthusiasm/return likelihood given half of them are automatic). As I speculated on Monday, it also appears that higher in-person/lower mail black voting is (mostly) offsetting what might otherwise be an immediate drop-off in black representation among the AIP crowd - though Silver showed how there's usually a surge in black voting early in the in-person cycle that quickly fades (historically, it has varied from cycle to cycle: sometimes black voters show up in the early days disproportionately, and sometimes at the very end - the former has tended to be the case in over the past 4 years).

Here's what 2020 GE non-voters are looking like rn. Obviously can't extrapolate too much from this.

https://twitter.com/lxeagle17/status/1339025414572806145

A good sign in terms of "meeting [normal] expectations": when you take explicitly-identified racial groups and reassign the "others" based on long-standing other-unknown registration discrepancies in GA, it's basically a stand-in for "young people who have recently registered/just started voting", with an approximate breakdown of 50% white, 40% black, 10% other; white number is on-point. Normally, the black figure would decline while the other figure would increase over time, but given non-black, non-white low-propensity behaviors, the former may hold at the expense of the latter.

EDIT: ugh, that's requests rather than votes. But still potentially great given that a disproportionately-young group had to manually request those ballots.

[Adam Griffin]

Does anybody know how many 18 year olds have “aged in” and are now eligible to vote in this election that weren’t in November?

The number I've seen floating around on social media is 23K but I don't know if that's been verified or anything.

Definitely in the ballpark. If you just breakdown population pyramids and calculate based on two months of growth without getting very precise, the real number could be anywhere from 20-30k.

However, only those who have either proactively registered to vote at 17.5 years or older (and prior to December 8 ) or those who visited a DMV to obtain their non-restricted licenses between October 6 and December 7 will actually be eligible to participate.

[Adam Griffin]

One good indicator in my opinion that isn't getting a lot of air time is the fact that the difference in racial composition between (accepted) ABM and AIP ballots is nearing zero. The difference between these two sets of votes remained substantially large throughout the entire AIP voting period (>5 points) in November and in 2018; it started comparably wide for the runoff but has shrunk quickly. Currently, the gulf as measured by white margin is only 2.2 points.  

This could be indicative of several factors. The most likely is that a large contingent of black voters who waited to vote until Election Day in November due to long lines are now voting early in-person instead.

I suppose the most pessimistic scenario for Democrats could be that a greater segment of white liberals who voted early in November have yet to cast their ballots in the runoff for whatever reason: as such, that could make the racial composition a lot less friendlier than would be expected.

But it's Georgia: when in doubt, bet on race. You'll never go broke in the long-term placing that bet over and over again.

[Adam Griffin]

Waiting until nearly 2 million Georgians have voted to dump and utilize this footage is stupid. If it were to have any impact, the time to do it would've been 10 days ago: at least then you would have the chance of scaring an optimal number of undecided white suburbrons.

[Adam Griffin]

I took a look at the non-GE voters - more specifically, where they are appearing disproportionately.



This is a mixed bag. There are multiple pros and cons for each side that I can see here, especially depending on whether there's a consistent lean in each jurisdiction for one party or another. So, I'm not too sure what can be said definitively about these voters at a granular/county level (obviously statewide, racial stats suggest this group is friendlier to Democrats than the GE electorate overall).

I'm not saying it's a correlation worth considering, but it reminds me a lot of November's swing map:

[Adam Griffin]

On the day before Election Day, Trump is coming to...Dalton. Sounds familiar:

The plotted stops of Harris, Obama and Trump:



Besides their presences in the final 48 hours of the campaign, where they are going says a lot as well. Particularly interesting to me is why Trump's team felt the need to choose Rome as opposed to, say, Gainesville or Canton.

(I saw some real internal data broken down by region at the beginning of October: let's just say that Rome/NW GA would make a better defensive choice for Trump if that data was accurate; it showed Trump sliding much more in NW GA than NE GA.)

[Adam Griffin]

I think anybody who predicts exact dynamics for how Election Day will look is a bit out of their league. The reality is that we have a series of variables that are historically strong coupled with a series of variables that are unique or one-time factors.

For instance, based on the share of 2020 GE voters who voted on Election Day (20%), it would likely be very difficult for the GOP to make up the deficit if those proportions held (they'd need to win 65% or more of ED vote). Still, one could also assume that Democrats doing better in early vote in the runoff compared to the GE would mean whatever ED vote exists will be even more GOP than in November (i.e. increased D cannibalization - a valid viewpoint especially given the uptick in black EV compared to November). Especially given the distrust the GOP sowed for mail voting and even early voting (muh Dominion), a 65%+ GOP ED electorate wouldn't be particularly shocking.

Still, runoff history suggests that a greater share of the vote is cast on Election Day than in standard general elections. In the 2018 runoffs, a whopping 69% (!) of voters cast ballots on Election Day. However, an unprecedented amount of cash and organizing has likely ameliorated this dynamic substantially (i.e. voters not being as aware of runoffs and/or not having to face long lines choosing to vote last-minute are behaving differently this cycle) and turned out way more people early than would otherwise turn out.

In short: we just don't know. As we hit the end of early voting, the raw vote is about 20% lower than it was in November; if you wanted to extrapolate trends from that, we're on track for an electorate of 4,000,000 voters. That'd imply somewhere between 1.2-1.5m ED votes (i.e. ~35% of total vote being ED), so make of that what you will.

[Adam Griffin]

And for those who want specifics from November:

Early Vote: Biden +5.73 (80.27% of vote)     52.35-46.62-1.03
EDay Vote: Trump +22.02 (19.73% of vote)  59.95-37.93-2.12

 

[Adam Griffin]

Dems lead down to +8.7 in my current model.

Again, based upon my methodology, I would lean towards this number slightly underestimating D's current lead, but that's speculation on my part.

Despite Republican's narrowing the margin, they aren't really netting any votes.

If in person votes broke how they did in the GE, the GOP would need an election day turnout of about 1.2 million to win according to my model.

Is that mathematically possible for the ED turnout to be that high?

And for those who want specifics from November:

Early Vote: Biden +5.73 (80.27% of vote)     52.35-46.62-1.03
EDay Vote: Trump +22.02 (19.73% of vote)  59.95-37.93-2.12

With the black share of EV being where it is relative to November, it lines up pretty well. Biden won EV in November by almost 6 points.

[Adam Griffin]

It's almost laughable how hard FoxNews is trying to completely smear Warnock- even saying that the reason Warnock does not have any events scheduled today is because he's hiding due to the "allegations that he ran over his wife's foot ...that came to light recently." (reality these have been out for a long time- and when examined, his wife's foot was perfectly fine).

In some ways it's not even a smart strategy for Fox.  If they are trying to keep the Senate in Republican hands, you would think they would target Ossoff... who polls indicate is likely to get fewer votes than Warnock.  I also think these over the top attacks on Warnock have only served to increase AA turnout.

Fox is even running this on their 6:00 News Hour.

In the words of opebo: "Warnock is a black, you see". If they can get their rabid audiences to care enough to turn out and vote against the scary black guy, they'll vote against the generic white guy as well. It's standard mobilization tactics: no need for persuasion here!

[Adam Griffin]

Given that E-Day was only 19% of the vote in the recent GE, I have a hard time believing 32% could be realistic. In the same vein, of course, the JMC poll today that had E-Day at only 7% of the vote was equally unrealistic. We should expect E-Day to be somewhere in the range of 15-25% of the total vote - any more or less is extraordinarily implausible.

See my post at the bottom of the previous page. 35% is perfectly possible. Even with unprecedented conditions, there's a very strong correlation between runoffs and a higher share of vote being ED. For it to be substantially less would imply an electorate of 3,000,000 voters which seems very, very unlikely (for a variety of reasons, not including most voters have 1-2 fewer days of early voting in the runoff when compared to the GE).

And for those who want specifics from November:

Early Vote: Biden +5.73 (80.27% of vote)     52.35-46.62-1.03
EDay Vote: Trump +22.02 (19.73% of vote)  59.95-37.93-2.12

 

Did the majority of Under 30s vote early or on election day in November?

This I do not know (I highly doubt a majority voted on Election Day). However, I am very confident that a disproportionate share of 18-29s voted on Election Day (i.e. maybe 30-35% compared to 20% overall).

One data-point I can provide is this: 3% of 18-29 voters in November cast their ballots by mail before in-person voting began (compared to 9% of voters overall).

[Adam Griffin]

Given that E-Day was only 19% of the vote in the recent GE, I have a hard time believing 32% could be realistic. In the same vein, of course, the JMC poll today that had E-Day at only 7% of the vote was equally unrealistic. We should expect E-Day to be somewhere in the range of 15-25% of the total vote - any more or less is extraordinarily implausible.

See my post at the bottom of the previous page. 35% is perfectly possible. Even with unprecedented conditions, there's a very strong correlation between runoffs and a higher share of vote being ED. For it to be substantially less would imply an electorate of 3,000,000 voters which seems very, very unlikely (for a variety of reasons, not including most voters have 1-2 fewer days of early voting in the runoff when compared to the GE).

And for those who want specifics from November:

Early Vote: Biden +5.73 (80.27% of vote)     52.35-46.62-1.03
EDay Vote: Trump +22.02 (19.73% of vote)  59.95-37.93-2.12

 

Did the majority of Under 30s vote early or on election day in November?

This I do not know (I highly doubt a majority voted on Election Day). However, I am very confident that a disproportionate share of 18-29s voted on Election Day (i.e. maybe 30-35% compared to 20% overall).

I think the real question is if the E-Day turnout is higher than 32%, will it less Republican than in the GE, and if it's below 32%, will it be more Republican. E-day turnout AND the way it breaks are the 2 variables here, and it's totally possible for Rs to win with only 25% e-day vote if it broke heavily in their favor and for Dems to win with a 35% e-day vote if it did not break super heavily GOP.

Also, would like to remind everyone that these current partisanship numbers are estimated and not concrete.

That's the gist of the question. A lot can probably be answered by an analysis of the black vote; how much of it is cannibalization from ED November voters turned off by the long lines they saw during the first few days of EV, how many are first-time/new voters and what percentages of D/R EVs from November have already early voted. I haven't delved too deeply into the data for the runoff, so those numbers may be available already.

[Adam Griffin]

I am sure this has been reported on but how will this Georgie vote count go in terms of what is processed first. Will this be similar to election night where it was red early and then later vote dumps were bluer and the question is if the republicans can max out on E-Day vote enough to hold on or is it the opposite?

"Smaller and more Republican counties -> larger and more Democratic counties" along with "Election Day vote -> early in-person vote -> mail vote" will be the general order of the evening. However and with 159 different counties, that doesn't mean all statewide in-person vote will be in before any mail vote, etc. Also very important: the city of Atlanta (5% of the state's vote and 10% of the Democratic vote) closes its polls one hour later than the rest of the state.

So yes, Georgia has a counting bias that favors Republicans early in the evening, with it being a case of whether or not Democrats can overcome those margins as more and more Democratic vote is counted (similar to VA). When there's 1 or 2% of the vote in, don't be surprised if the GOP leads 65-35 or whatever.  

2020 did have a break with historical trends - at least in the early stages of vote count:

[Adam Griffin]

Ossoff said on "The Last Word" that young voters have shattered voting participation records, etc.  He said a list of several demographics that have shattered records-

Did he just get carried away by including young voters... or has there really been strong turnout among young voters (I thought I saw somewhere that youth vote was down a fair amount)?

Among early voters in November, the 18-29 demographic was somewhere in the 10.x% range; as of Tuesday's vote, they were at 11%, so that's likely accurate.

[Adam Griffin]

Why are Hispanic registered voters in GA such lazy-ass voters ?

Their turnout so far is 20%, compared to 40% each for Whites and Blacks and 35% for Asians.

Latinos are disproportionately young and young people disproportionately vote on Election Day. Still, being somebody who organizes in the most-Latino county in the state, a great deal of it comes down to a lack of concern/attention paid when it's not a presidential contest (again, young people). We saw huge drop-offs in the 2018 runoff and the broader trend appears nationally in every contest.

[Adam Griffin]

https://twitter.com/Nate_Cohn/status/1345120553519222785

[Adam Griffin]

Has Deborah Jackson been out campaigning for Warnock? She did get nearly 7% in the original race..

Deborah Jackson got 7% because she was the first Democrat on the ballot in a 20-person race. She has no legitimate pull with the vast, vast, vast majority of voters who preferred her.

[Adam Griffin]

Trump won this county 70-29

https://twitter.com/ElectProject/status/1346125171132915712

And it's literally the 2 most GOP precincts that are closed; only about 10% of the county's November vote, though:

[Adam Griffin]

On a scale of 1-10, how laughable is the bold prediction I'm about to make?

Both dems win, but Ossoff overperforms Warnock.

Warnock will almost certainly do better in the metro areas, while history would suggest Ossoff will probably do better everywhere else (people still don't grasp that being a white Democrat in GA benefits rural margins compared to being a black Democrat: yes, even in 2020).

The big question in such a scenario is if/how much more Ossoff does better in rural areas relative to Warnock's performance in metro areas; obviously the metro areas are a much larger share of the state (~70%), but if Warnock only does 1-2 points better there while Ossoff does 3-4 points better or more in rural areas, then Ossoff outperforming Warnock is possible.

[Adam Griffin]

The frustrating thing about all of this campaign mail and phone/text barrage is just the sheer amount of waste that is occurring.

To put it into perspective: more money has poured into GA for these runoffs than the Obama campaign spent nationally on its entire 2012 re-election bid. Let that sink in for a second.

I voted over a month ago. Within a few days, all of the glossy Warnock & Ossoff mailers (from the campaigns themselves) ceased. Thank God they at least have some common sense.

However, all of these PACs, outside groups and even the state party continue to pepper me with mail. Anybody with a voter file (and all of these groups certainly have them) can tell within 24-48 hours that a particular individual has voted.

I get that there is a certain lag for setting up mail campaigns - and even more a lag given USPS holiday delays compounding the situation - but this is inexcusable. Within 10 days max of voting (and that's with the aforementioned USPS lag), no individual should still be receiving campaign literature in the mail.

It's obvious that a bunch of groups spending and allocating resources here either have no idea what they're doing or are just so flush with cash that they can't be bothered with filtering their voter file universes on a daily basis to exclude people who have already voted. There's probably more financial and logistical waste occurring in these runoffs than occurred across all of the competitive Senate races combined in 2020. And you know what? I can't complain in every facet: my county party has raised basically as much via ActBlue over the past 6 weeks from random out-of-state contributors as we would raise from local donors in an average year.

Somebody who voted 5 weeks ago shouldn't be getting 10 mailers, 10 texts and 5 phone calls per day telling them to vote.

[Adam Griffin]

Supposedly there are more than 20k people waiting for Trump's rally at the Dalton airport. Ugh...we already have the highest COVID rates in the state and this is just gonna add fuel to the fire.

[Adam Griffin]

Wrt above discussion:

I am sure this has been reported on but how will this Georgie vote count go in terms of what is processed first. Will this be similar to election night where it was red early and then later vote dumps were bluer and the question is if the republicans can max out on E-Day vote enough to hold on or is it the opposite?

"Smaller and more Republican counties -> larger and more Democratic counties" along with "Election Day vote -> early in-person vote -> mail vote" will be the general order of the evening. However and with 159 different counties, that doesn't mean all statewide in-person vote will be in before any mail vote, etc. Also very important: the city of Atlanta (5% of the state's vote and 10% of the Democratic vote) closes its polls one hour later than the rest of the state.

So yes, Georgia has a counting bias that favors Republicans early in the evening, with it being a case of whether or not Democrats can overcome those margins as more and more Democratic vote is counted (similar to VA). When there's 1 or 2% of the vote in, don't be surprised if the GOP leads 65-35 or whatever.  

2020 did have a break with historical trends - at least in the early stages of vote count:

[Adam Griffin]

My dad was going on today about how black farmers in South Georgia are going to make Perdue win handily.

I'll bite.  Why?

Maybe his dad saw this tweet

https://twitter.com/ewerickson/status/1345889577899356162?s=21

Lol I was trying to find the source of this and you found it. IIRC Erickson is his favorite pundit. Lmao thanks.

Honestly there is some truth to it. Rural black voters in South GA (particularly south of Columbus and west of I-75) have time and time again shown a propensity pre-2020 to support various Republicans at 15-20%. While this tends to happen downballot more so (think state representatives and local elected officials), the Perdue name itself has some gravitas (farmers or not). Sonny got fairly crazy numbers of black support in his re-election (>15%). It's pretty obvious from statewide analysis that Trump did in fact get no less than 10% of the black vote: I would not be surprised (especially given the swings we saw in predominantly black rural counties) if he got closer to 15% or more in rural South GA.

Anybody can take a look at this tool and see various SWGA precincts where Perdue did better than Trump by a gradient - and the gradients are 20-point intervals; the effect would certainly be more noticeable if this map displayed margins by 5-point gradients instead. I really want to see a swing map by precinct of Georgia.

https://gisanddata.maps.arcgis.com/apps/MapSeries/index.html?appid=5d930a3281ad403391b1d9b3f7f9d145

[Adam Griffin]

Here, I made this a few weeks ago. You can see that corner of SW GA is even redder than Atlanta suburbs.

This is an interesting map (to me, at least) if for only one reason: it reminds everybody that NWGA is basically the only part of the state where poor white people still exist en masse (as a percentage of total pop). NEGA is increasingly being filled with rich out-of-state whites, metro and urban areas are of course very diverse in every facet, and rural South GA is a big mixture of poor white and poor black communities. It's basically the "purest" region when balancing base socioeconomic variables. Of course, it does also have the largest Latino population of any broader area outside of the Gwinnett/Hall region, which creates a unique variable (that is often minimized in runoff/low-turnout contests).

Combined with the media market segmentation (Chattanooga predominantly serving the 7 most northwestern counties), it's pretty easy to see why Trump outperformed Perdue in a largely poor, white, manufacturing-based region. Sometimes it helps to imagine NWGA as a non-unionized, Southron analog of the Rust Belt.

It's also going to be a good indicator tomorrow I think in terms of whether there'll be enough ED vote to give Republicans a win. The same area is also seeing some of the lowest turnout rates thus far compared to November; it's obvious the Republicans are worried about this and it likely explains why Trump decided to close out the runoff with a rally in Dalton (and close out the GE with a rally in Rome).

Keep an eye on fully reporting precincts in non-Paulding CD-14 tomorrow and compare overall raw totals and ED raw totals to November. They may be the earliest single-best indicator of what to expect.