GA-6 Special election discussion thread

[Adam Griffin]

I'm hearing that DCCC isn't likely to get involved unless this gets whittled down to one declared Democratic candidate in the very near future. The run-off is a guarantee, but don't expect DCCC support prior to that with multiple Democratic candidates; it'll speak volumes about the organizational efficacy of local and state Democrats in the area to unite behind a candidate or not, and that will affect DCCC's decision. If they can't get involved before the run-off, they likely won't get involved after the fact.


The biggest challenge with this district is "The Discrepancy". It has existed since at least 2008, and likely even before that. These voters are exactly the types who are amenable to a Democratic presidential candidate but do not want local Democratic representatives or even House/Senate types. I've always maintained that for local races, it makes sense: these are exactly the types of people you'd stereotype as being NIMBYs who want a socially-responsible figurehead as President, but "muh low [property] taxes" mean they vote GOP in all other races.

There are dozens of precincts in the district and the adjacent areas where Obama's margin outperformed the aggregate D's by 10-20 points in 2008; some where Obama won by nearly 20 points but the average Republican carried by high single-digits or more. Many of these precincts were among some of the biggest swingers in GA to Romney in 2012. This was also on display in 2016:

Stooksbury got 38% (D was 35% in 2012) and Clinton got 47% (Obama was 37% in 2012). Only about one-third of the Clinton improvement "flowed" down-ballot. I've put that in quotes for a reason. When you look at the demographic shifts in the area, you basically walk away with the understanding that the only thing that flowed downballot was demographic shift; one-third of Clinton's improvement was attributable to demography and two-thirds was attributable to independents and Republicans who crossed party lines to vote for her. Virtually none of the Romney-Clinton voters defected in any other race.

You then have to contend with the fact that this is a special election. That's obviously going to give the Republicans a slight advantage/margin improvement right off the bat in any situation. Can that single disadvantage be reduced or eliminated? Sure - but it just means that it's going to take that much more work to short-circuit the natural dynamic of the race. This is also one of the wealthiest (the wealthiest?) CD in Georgia, which means that turnout for a special will be relatively high. That might sound good, but when you're working as a Democrat against the natural lean of a district, you're probably going to want the baseline turnout in a special to be as low as possible, so that any set number of pro-Dem voters that you do mobilize have a larger impact on the margin.

All in all, I don't see Ossoff being able to do much better than 40-42% - and that's assuming he runs a great campaign.

[Adam Griffin]

I could see them pulling it off if they ran exclusively on "slow this chaos down!" Or "let's send a message!" and not as an ideological liberal. That'd be similar to how Scott Brown pulled off his win

Ossoff is just not the right candidate to pull off a upset he a C list recruit at best. Holcomb was the one to to potentially make this a race. If we had Holcomb in GA-6 and Schweitzer running in MT-AL that would show national Ds are serious. So far recruiting shows their saving their energy for 2018.

Oh, from what I've seen from Ossoff so far, I'm not impressed. He's just running a one-note "I hate Trump" campaign it seems like, and he comes across as an ideologue to me. Dems need a Holcomb or a Gottheimer-like pol for this district. Not some 30 year old filmmaker.

To be fair, Taylor Bennett ran in the 2015 special for HD 80 (Romney got 55% there in '12; district is entirely within GA-6) on a one-note campaign of, effectively, "I oppose these bigoted 'religious freedom' laws", and won 55-45.

[Adam Griffin]

By the way, the answer to this is still "no".

[Adam Griffin]

LMAO who is dumb enough to use Trafalgar? Also, no way Trump is +10 in a district barely to the right of the national popular vote.

[Adam Griffin]

Day 5: https://mobile.twitter.com/ElectProject/status/848008061709168640

Ballots: 8,110

McDonald's numbers (not Cohn's):

• Dem 3591 (44%)
• Unknown 2639 (33%)
• Rep 1880 (23%)

(Was 45% D, 32% U, 23% R yesterday)

[Adam Griffin]

Is that total, or just for Day 5?

Total

[Adam Griffin]

That's a big jump among nonpartisans. Wherever they lean will decide this race.

I think you may be comparing Cohn's numbers to McDonald's, which are quite a bit different. The ones I posted are McDonald's; McDonald's numbers yesterday were 45 D/32 U/23 R. For comparison, Cohn's numbers from yesterday were 55 D/16 U/29 R.

[Adam Griffin]

How many people do we expect to vote in the primary?  Is getting over 200 thousand too out there?

Early turnout as a raw number is on par with 2014's early turnout, for what it's worth. There were 210k votes cast in the GA-6 race overall in 2014.

It's the wealthiest CD in GA and one of the wealthiest in the South, so I expect overall turnout for a special election will be much higher than average - even before factoring in just how competitive this race is now apparently becoming.

[Adam Griffin]

I think you may be comparing Cohn's numbers to McDonald's, which are quite a bit different. The ones I posted are McDonald's; McDonald's numbers yesterday were 45 D/32 U/23 R. For comparison, Cohn's numbers from yesterday were 55 D/16 U/29 R.

Just curious - but where are they getting those numbers from?

Voter file data. From the looks of it, McDonald is playing it safe, only relying on those with well-established histories of voting in party primaries, while Cohn is likely including everybody who has voted in a primary - even if they have voted in just one - and may also be using voter file scoring models to further predict how a greater share of the "unknowns" is likely to lean.

This is arguably why McDonald's share of unknowns is in the mid-30s, while Cohn's is in the low-20s. Honestly, I'd bank on Cohn's being the more accurate of the two: McDonald has acknowledged in the past that his access to voter file data is often less accurate/older than some others; on the other hand, I imagine Cohn has access to a premiere voter file solution like Catalyst.

[Adam Griffin]

11 Alive/SUSA Poll:

Ossoff 43
Handel 15
Gray 14
Moody 7
All Others 15
Undecided 7

[Adam Griffin]

^^^ Better hope that all of those outstanding/likely Ossoff voters are pretty young...

You must be logged in to read this quote.

[Adam Griffin]

Does this come up for anyone else when they type Jon Ossoff in Google? It seems weird that would be the top search about him.

http://i.imgur.com/gS4Jkhz.png

!!!

Search engine manipulation!

[Adam Griffin]

AKA why the GOP tries to undermine early voting:

[Adam Griffin]

R turnout nearly equaled D turnout. The best day yet for Rs; ridiculously bad for this district, but if this trend continues it will make it hard for Ossoff to hit 50.

Even at a 42D-40R split like yesterday, there's still a decent chance Ossoff got 50% of those votes.

Yeah but every recent election has shown that election day votes are far more republican than early votes, ossoff needs to be winning early votes by a much larger margin than that to get 50%

Overall EV is 52 D/31 R/17 U according to Cohn, which - assuming the huge spike in unknown, first-time and newly-registered voters leans somewhat toward Ossoff - means he's above 60% of the early vote. If EV percentage for him holds at 60% and EV ends up being only 25% of votes cast, the aggregate Election Day GOP percentage could mirror that (60-40) and Ossoff would still be at or above 45%.

[Adam Griffin]

[Adam Griffin]

Based on the 2% Trump win in Georgia's 6th coupled with the strong wave of Democratic enthusiasm why isn't it at least likely that Ossoff reaches 51% on Tuesday? So far the Democrats are running ahead of their 2016 numbers and this is a suburban Atlanta district. Why would they vote for a Republican if they believe they're voting on the Trump Administration?

I don't have the side-by-side numbers to definitely confirm it, but I'm expecting that a greater share of likely voting Democrats in GA-6 have already early voted compared to who had done so in KS-4 at this comparable point. I remember Cohn saying a few days ago that something like 80% of the people remaining who had a 80% chance or greater of voting in a midterm were Republican or Republican leaners.

That 30-point gap in preferences in Sedgwick between EV (Thompson +22) & ED (Estes +10) could look quite small compared to what we see in GA-6 if there isn't substantial breakdown of usual voting fundamentals.

[Adam Griffin]

Based on the 2% Trump win in Georgia's 6th coupled with the strong wave of Democratic enthusiasm why isn't it at least likely that Ossoff reaches 51% on Tuesday? So far the Democrats are running ahead of their 2016 numbers and this is a suburban Atlanta district. Why would they vote for a Republican if they believe they're voting on the Trump Administration?

I don't have the side-by-side numbers to definitely confirm it, but I'm expecting that a greater share of likely voting Democrats in GA-6 have already early voted compared to who had done so in KS-4 at this comparable point. I remember Cohn saying a few days ago that something like 80% of the people remaining who had a 80% chance or greater of voting in a midterm were Republican or Republican leaners.

That 30-point gap in preferences in Sedgwick between EV (Thompson +22) & ED (Estes +10) could look quite small compared to what we see in GA-6 if there isn't substantial breakdown of usual voting fundamentals.

Will it matter if Ossoff is at 60% in the early vote? According to Cohn, the current EV electorate is 46% D, 37% R, 17% NA, which makes that quite difficult, but perhaps there are a fair amount of "crossover" votes. What % of the overall vote do you expect to come from the EV?

In KS-4/Sedgwick County, Thompson got 61% of the EV; he barely carried Sedgwick overall and lost the CD by 6-7 with Sedgwick comprising 70% of the CD's vote. If the gap between ED/EV is even modestly larger than that for GA-6 as a whole, then there's no way he hits 50% or comes even reasonably close.

[Adam Griffin]

What % of the overall vote do you expect to come from the EV?

I think it'll be a larger share than KS-4...around 40% - maybe as much as 50%. In regular elections, it has recently not been uncommon for a nominal majority of votes statewide to be cast before Election Day. A district like this socioeconomically + less worry about longer lines on ED will probably mean it's less than that, though. We may see another 20-30k votes come in over the next 3 days. There were 7,000 votes cast today alone (out of 39,000 total).

EDIT: One potential variable working in the opposite direction here is Dekalb. There are still no well-positioned early voting locations in the more Democratic portions of the county, so ED turnout here will likely be disproportionately larger than its share of the electorate on Election Day.

[Adam Griffin]

So you feel the Democrats are cannibalizing too many early votes? I'm not sure what you're saying, although that's a pretty interesting assessment.

RRH polled the race; there's indicators that Ossoff might be stronger than we anticipate. Just a thought.

If this is correct, then yes:

https://twitter.com/Nate_Cohn/status/851623999436197889

[Adam Griffin]

Let's assume things play out like they did in KS-4, but with a slightly higher percentage of voters casting early ballots. If 40% of the vote is EV and plays out along the margin above, then the remaining 60% would be Trump +14. This would produce an electorate that Trump won by one point (identical to 2016 electorate). If these GOP early voters are slightly more D-leaning, then we'd end up with a plurality/majority-Clinton electorate, but not by much.

Since the EV is likely to continue its shift toward the GOP, it's more likely that the electorate winds up being one that Trump won by a few points (unless, again, these GOP voters are substantially more pro-Clinton/Ossoff than we might expect otherwise).

I'm personally not confident that Ossoff can win unless the electorate as a whole is something like Clinton +6 or greater, because "Trump =/= Generic GOP" and many of these GOP/Clinton voters have sane choices to consider.

[Adam Griffin]

[Adam Griffin]

Remember that if Ossoff doesn't make 50%, there's still the runoff, and the expectations game will probably evolve quite a bit between now and then.

Ossoff also has a huge warchest where he could immediately start bombarding the Republican winner with negative ads the instant the race is called next Tuesday.

I'm not so sure. His burn rate has been very high - almost comparable to Ben Carson-type fundraising schemes - and only had like one quarter of the $8 million raised on hand as of a week ago or so. That's still probably a lot more than any one GOP candidate will have at the onset, but I imagine the discrepancy will narrow very quickly once there's a nominee...and even more so if Ossoff's massive fundraising numbers aren't really indicative of low-cost, high-yield fundraising.

[Adam Griffin]

Not good, but definitely more in line with what I was expecting before all of this hoop-la made me think there might be a shot for Ossoff (my original prediction was that Ossoff would get 41-43% of the vote; I haven't necessarily abandoned that belief). This poll seems to be GOP-friendly but the fact that it shows Ossoff only at 50 among early voters - presumably measured before the past couple of days of heavy GOP voting - is a very bad sign.

[Adam Griffin]

This was always bound to happen. It's a 60/40 Republican district. I'd rather they get as much of their vote share counted prior to ED as possible than have a 40-point difference between EV/ED and get our hopes up on Election Night when the early vote totals get counted first.

[Adam Griffin]

Yet the early vote is still much older, whiter, and more male than in 2016...

Also important to note that the most nonwhite part of the district never had a good early voting location.

The southern portion of the district in Dekalb isn't really much whiter in CVAP terms (if at all) than the section of north Fulton that straddles 400. The white share of total population in that part of Fulton is about 10 points higher, but the share of the total population in the southern areas of Dekalb-6 is nearly 30% Latino (around one-third of that in the Fulton part). These two areas have roughly the same total population. Black population share is identical, too.



Still, turnout isn't very strong in either of those areas, despite the north Fulton area having ample access to EV sites. It's higher in Fulton - and yes, that's definitely partly because of EV site access - but nothing to write home about considering it's right in the middle of the greatest cluster of EV sites. I'm not very confident that ground will be made up on Election Day in Dekalb, in large part because of Latino voters.