GA-SEN 2022 Megathread: Werewolves and Vampires

[Florida Man for Crime]

For everyone's reference, here is the demographics of 2022 Electorate in Georgia. Quite encouraging that Warnock won by 1% in an electorate which was only 26% Black,  which is much lower than 2016, 2018, 2020.

https://twitter.com/WinWithJMC/status/1597947327763079168

The reason why the "black" vote share has tended to go down in Georgia over he past decade or so is, as Adam Griffen alluded to a few posts back, the fact that many of the newer voter registrations don't have race data. So (particularly if he "White" share is also down) what it means when the Black share is down is that there are more newer registrations voting which don't have race data. By the way, those same people are also disproportionately younger and more concentrated in the Atlanta metro.

So a lower "Black" share is not necessarily always bad for Ds and good for Rs.

[Florida Man for Crime]

Updated figure on % of GE EV by district, now with the right IV/DV setup and polynomial fit since the relationship is actually fairly messy:

I would have about 0.1% confidence in that line (and sounds like you would also not have much more since you mention it is pretty messy). It also has the shape it has because of the assumption that it has a polynomial functional form. You would get very different results if you assumed even a different degree polynomial.

Part of the issue is there are not many data points, so if it were possible to do the same thing for state house districts or something like that which is a bit more granular, you might be able to see some more reliable patterns. Could be interesting to see possibly.

[Florida Man for Crime]

Unless ED turnout is close to the size of EV turnout (or greater), Walker is cooked.

Do you think it is not actually a real possibility that election day turnout may be higher than expected? I would think that with the shortened early vote period there are probably some people who may have voted early in the past during the longer periods in previous elections might end up not voting on election day because they are busy, because they forget, or even because of some freak weather or something.

And it is plausible those sorts of people could be disproportionately D voters. That is the one thing giving me pause, the possibility that Rs could win by being successful enough at suppressing the D early vote that they could end up winning on election day.

The early vote data on its face certainly does look very good, but with that major caveat regarding the shortened period.

[Florida Man for Crime]

Wait if people who intended to vote early stay home on Election Day, wouldn’t that mean Election Day turnout is lower?

I’m not sure if you mean that high or low turnout is good for Walker?

Depends on what "higher than expected" entails. My only true expectation personally is that runoff turnout will not exceed general election turnout. We're on track for 1.92m EVs when all is said and done, so 2.02m is the absolute ceiling in any somewhat plausible scenario. However, is it that realistic to expect even a number close to that?

Maybe I'm touching on part of what you mean here (along with TG, I'm a bit confused), but if anything, a greater share than usual of those who are waiting to vote on Election Day may be metro voters who were turned off by huge wait times during the EV period that we haven't seen since the early Obama years.

Sorry for the lack of clarity. I'll try to make it clearer.

What I was worried about is the possibility that the shorter EV period has the effect of selectively reducing turnout among "the sort of people who typically in the past have voted early" because their normal method of voting has been limited but not among "the sort of people who typically in the past have voted on election day," because their normal method of voting has not been limited. And since past early voters have tended D and election day voters have tended R, that could have the effect of helping Walker if that turns out to occur.

I will make up an example scenario below using some numbers to make that more concrete (looking up numbers from Georgia SOS which I did not bother to do before).

I fully expect the EV/ED gap to be larger than in recent general elections (in 2020 & 2022, it was 26-28 points; in the 2021 runoff, it was 40 points). The problem for Walker is voter file analysis suggests that Warnock is leading among EVs by 15-20 points right now (in contrast, he won EV in the 2021 runoff by 14 points).

I checked with the individual I referenced prior: reverse-engineering their breakdowns, their poll is projecting 3.3m voters (which points toward an ED total of 1.4m voters). The EV/ED candidate support gap in that scenario would need to be close to 50 points for Walker to have a shot.

Loeffler got 63% in the 2021 runoff among an ED electorate of 1.3m voters (with the total electorate being 4.5m voters!). Walker not only needs Election Day to have more raw voters than we saw in an almost certainly higher turnout contest (2021 runoff), but he needs to win them by more than Loeffler did in order to win.[/quote]



What I am seeing from the detail file here https://results.enr.clarityelections.com/GA/115465/web.307039/#/detail/10100

is that in the 2022 general, results were:

EDay - Walker 794384 (56.3%), Warnock 575873 (40.8%)............ Walker + 15.5% margin
Absentee - Walker 74845 (30.4%), Warnock 166257 (67.6%)............ Warnock + 37.2% margin
Advance - Walker 1038220 (45.6%), Warnock 1202113 (52.8%)............ Warnock + 7.2% margin
Provisional - (won't bother since the numbers are small)

If you combine Eday + provisional into one election day category and Absentee and Advance into one early category, then you have:

EDay (incl prov) - Walker 795377 (56.3%), Warnock 577747 (40.9%)............ Walker + 15.4% margin
Early - Walker 1113065 (44.1%), Warnock 1368370 (54.3%)............ Warnock + 10.1% margin


According to he numbers you posted earlier, 1,852,576 people have voted already in the 2022 runoff (Absentee + Advance).

And let's assume, based on the estimate you posted which sounds like it is probably based on voter file support scores, that Warnock will win the early (Absentee + Advance) vote by 17.5%. If those voter file support scores, when applied to the list of people who voted early (absentee + advance) in the 2022 general matches the actual results of the 2022 early vote and gives Warnock around a ~10.1% margin (do you know if it in fact does?), then those support scores are probably pretty accurate now for the runoff early voters as well. If that does not match, I would be more wary.

But anyway, if we assume 17.5% margin from the early votes, that would put the 2022 runoff early vote at an estimated 1,088,388 Warnock to 764,188 Walker, or a 324,200 vote margin for Warnock.


And from here - https://results.enr.clarityelections.com/GA/107556/web.274956/#/detail/10100

In the Warnock-Loeffler 2020 runoff, 1,316,760 votes were cast on election day, and Loeffler won them 830,949 to 485,811, or a 345,138 vote margin.


So now let's suppose for my example that basically everyone who voted on election day for the 2020 runoff now shows up for the 2022 runoff as well, and if they voted for Warnock then, they vote for him again, and if they voted for Loeffler then, now they vote Walker. So that gives us 485,811 for Warnock and 830,949 for Walker, same exact thing as the Warnock-Loeffler runoff election day.

And in addition to that, let's suppose that there are also 100,000 other people that vote on election day, which are people who normally would have voted early in previous years, but did not do so this year because of the shortened early vote period. Since these are "the sorts of people who normally vote early" they are a good group for Warnock, and let's say they also vote for him by 17.5%, just like the early voters, so that would mean they break down 58750 for Warnock and 41250 for Walker.

In total, these assumptions give us:

Warnock: 1088388 + 58750 + 485811 (Early + People_who_usually_vote_early_but_voting_election_day_this_time + Election Day)

Walker: 764,188 + 41250 + 830949 (Early + People_who_usually_vote_early_but_voting_election_day_this_time + Election Day)

Or...
Warnock: 1088388 + 58750 + 485811 = 1,632,949
Walker: 764,188 + 41250 + 830949 = 1,636,387


So despite the early vote looking seemingly good for the Dems, those (entirely hypothetical) numbers actually end up adding up to a narrow Walker win.

This is the sort of scenario I am worried could still be possible, and I don't think the particular assumptions we are making here are really unreasonable (and also are not particularly favorable for Rs in a lot of respects, since we are assuming Warnock wins the early vote by a larger percentage than he did in the 2020 runoff etc).

At any rate, I'm expecting somewhere between 1.2m and 1.4m ED votes, as this has been the case in the past 2 major high-turnout contests. That ED share will work out to somewhere between 36-42% of all votes cast.

I would also note that this scenario above has 1,416,760 election day votes, which is the same as the Loeffler-Warnock runoff election day turnout plus an extra 100,000, and which ALSO is basically within your own 1.2m to 1.4m expected election day vote range.

So I could still see a path to victory for Walker along these sorts of lines, and hypothetically he could do better also if e.g. Warnock "only" wins the early vote by 16% or 15% or something rather than by 17.5%.

[Florida Man for Crime]

And the reason for the numbers in my hypothetical scenario adding up to a narrow Walker win is really just that early vote turnout was suppressed by the shorter early vote period, and although we are assuming that 100k people who normally would have voted early do end up voting on election day instead, there are also a lot of people who normally would have voted early (and lean Dem) who end up not voting on election day, because they are not used to voting on election day. If there were a longer early vote period, a lot of them would have probably voted early, but this is a hypothetical example where voter suppression works to create an R victory.

[Florida Man for Crime]

Walker not only needs Election Day to have more raw voters than we saw in an almost certainly higher turnout contest (2021 runoff), but he needs to win them by more than Loeffler did in order to win.

And finally, in my hypothetical scenario, Walker does get higher raw election day turnout than in the 2020-21 runoff, but there is a logical explanation for why that could plausibly be the case, namely the shortened early vote period this time.

And in contrast to your statement that Walker would need to win them by more than Loeffler did to win, he wins them by a LOWER margin than Loeffler, but nevertheless that is enough for a narrow Walker victory.


So while I would rather be Warnock than Walker at this point, I don't think we are necessarily quite out of the woods yet.

[Florida Man for Crime]

Think of the numbers a Trump rally would have brought in! But this guy, Walker, was very ungrateful and asked him not to come. Sad!

If Walker loses (and if he does worse than in the general), it will be proof that it is because he did not have an in person in-state Trump rally. Rs will see that not supporting Trump and not using his help doesn't work.

[Florida Man for Crime]

Warnock will win by 10+ Points. The Question is when will the Race be called. My guess between 8 and 9pm ET!

This is the only time I can ever recall seeing you make a prediction that is more Dem friendly than the conventional wisdom.

Why now?

Is it because you have dumped Trump for DeSantis and you figure if Walker loses badly it will reflect poorly on Trump and ultimately help DeSantis?

[Florida Man for Crime]

The best was that annoying Republican (not Josh Holmes but something similar) who is always on their panels saying something like "after this midterm, voters clearly want new leaders and new blood" trying to equate Biden and Trump and how both should step aside in 2024, as if Biden did not just deliver an incredible midterm performance! Absolute insanity.

You can be positive about Biden and support him etc if you like him, but let's get real here - Biden did not in fact deliver an incredible midterm performance. By that, I don't mean that Dems didn't do well in the midterms (and I agree that at least under the circumstances and relative to expectations it would be fair to call it "incredible"), but it just means that Biden was not personally responsible for it, and it is silly and unfounded to attribute it to him personally.

Biden had subpar approval ratings in the 40s of the sort which in the past co-occured with large scale losses. If Biden actually had positive approvals, you could maybe make some sort of argument, but that was not the case in exit polls (exit polls are not perfectly reliable, but they are reliable enough for this broad statement). Exit polls also had a supermajority of voters that said they didn't think that Biden should run again. Voters may or may not have been right about that and obviously it is up to Biden, not them, whether he runs, but undeniably that is what midterm voters thought.

The reasons why Dems did relatively well in the midterms were not Biden, but primarily Trump and Dobbs.

If you really want to give Biden credit for something, I suppose you could correctly say that he did not sign any major, important, and controversial legislation like Obamacare or something which could have alienated and fired up some voters in opposition to him and in opposition to Democrats. However, it would be a pretty bizarre thing to give Biden "credit" for, because it is not what he wanted to do, but instead was forced on him by the fact that Dems had only a 51-50 advantage in the Senate which was subject to Manchin and Sinema, which made it impossible to get anything significant passed which truly changed the trajectory the country was on or addressed and decisively dealt with a major issue.

[Florida Man for Crime]

Biden's approval rating or asking if people want him to run again are not really relevant here.

It is certainly 120% relevant to whether he is responsible for Dems doing relatively well in the midterms lmao. Bill Clinton in 1998 for instance had good approvals when Dems did relatively well in the 1998 midterms, so you can much more plausibly say that he helped.

Biden was the one who refused to back down in making election integrity and fighting for democracy, along with abortion rights, a central theme to the midterm elections. He literally did a massive speech on it the week before the election. Turns out, it was incredibly prescient. You can't just remove him from this and say, hey, you oversaw this election and determined a lot of the ultimate messaging, but your approval was only 44% so we're not giving you any credit.

The way you are talking about it, it sounds like you are trying to imply Trump's coup attempt and Dobbs were not issues until Biden mentioned them, and that people cared about them only because Biden mentioned them.

I don't know your age etc, but in case you are too young to remember much from previous Presidents, I can assure you that Obama for instance made plenty of "massive speeches" before the 2010 and 2014 elections, which barely moved the needle. This was not because Obama was a bad speaker in comparison to Biden, a bad President, or that he was not talking about important issues that people cared about. Rather it is because Presidents are not omnipotent demigods and are not, in general, responsible for everything that happens in the country.

Your argument is just as silly - and for precisely the same reasons - as Republicans who like to pretend that Biden is singlehandedly responsible for global inflation.

To try and act like he is on the same level as Trump is ludicrous.

Honestly, what are you talking about??? I never said anything of the sort.

Biden has objectively been an incredibly strong first time president, but he just should not run again, when nearly every other first termer has because.... reasons? That line of thinking is ridiculous.

Whether Biden should or should not run again and what reasons he may or may not have to do so is utterly irrelevant to the question of whether midterm voters thought he should run again. When asked the question, a clear supermajority said no. You can spin that however you want, but it is a simple fact. If midterm voters were voting D because they liked Biden so much, it stands to reason that fewer of them would have said that.

And incidentally, there are also other similar/related questions from the exit polls which also do not support your argument at all, like this:

https://www.cnn.com/election/2022/exit-polls/national-results/house/0



You might also find it interesting to compare to e.g. 2014:

https://www.nbcnews.com/politics/elections/2014/us/house/exitpoll/



hmmm those look awfully similar, don't they? Maybe it will make you think a bit.