Early Voting thread.

[Unelectable Bystander]

Total Early Votes: 237,512
Mail Ballots Requested: 8,307,046

Party           Count Percent
Democrat   60,146 57.9
Republican   26,792 25.8
None/Minor 16,961 16.3
TOTAL 103,899 100.0

North Carolina, the state with the most votes that has early statistics, is showing numbers roughly in line with 2020, if a little better for Democrats.

Obviously still early but I would call this a good sign for Budd if the electorate is very similar to that of 2020 (not that party registration proves that it is).

1) this would mean that possibility of a 2018 style absolute democrat turnout advantage blitz seen in the special elections is less likely

2) As skill and chance pointed out, it seems that the people who vote a month early are trending hard away from republicans

3) It’s hard to imagine Budd doing so much worse than Trump with independents that he loses in a neutral turnout scenario

[Unelectable Bystander]

https://twitter.com/tbonier/status/1579110131887136771
Good thread here that features a discussion of early voting.  Click to read the whole thing.

This seems like fair analysis for the most part. Still, nobody ever goes into an election publicly expecting a shellacking. The fact that someone like Bonier would cite any evidence for a favorable republican night makes me more confident in a R+2 ish prediction. For example, even though 2018 had good polls and 2020 had a trifecta, there were a lot of democrat expectations that turned out to be insanely unrealistic after the fact.

[Unelectable Bystander]

https://mobile.twitter.com/WinWithJMC/status/1580889276468064257

How is this special midterm looking?

There’s one caveat here and that’s the fact that the covid caucus might now be willing to vote in person which could diminish democrats’ actual mail-in numbers. But if the total mail-in numbers start to approach 2020 numbers or if republicans approach their own totals for 2020 mail-ins and these patterns continue, I would call it.

We’ve also seen that the early/E-day disparity is wide as ever in the primaries so I don’t think democrats doing a 180 on vote method can explain all of it.

[Unelectable Bystander]

https://mobile.twitter.com/WinWithJMC/status/1580889276468064257

How is this special midterm looking?

There’s one caveat here and that’s the fact that the covid caucus might now be willing to vote in person which could diminish democrats’ actual mail-in numbers. But if the total mail-in numbers start to approach 2020 numbers or if republicans approach their own totals for 2020 mail-ins and these patterns continue, I would call it.

We’ve also seen that the early/E-day disparity is wide as ever in the primaries so I don’t think democrats doing a 180 on vote method can explain all of it.

Can we though?

High propensity democrats are likely to vote early and by mail. Low propensity democrats would likely vote on election day? If enthusiasm is as high as polls are showing, there's probably alot of democrats who are just going to vote on election day. Myself included.

Quinnipiac asked this question in a recent poll. It’s just one data point but they found:

Warnock winning mail voters 66-28
Warnock winning early in person 59-40
Walker winning E day 60-37

Thanks to 13% of democrats (vs 7% of republicans) voting by mail, and 42% of republicans (vs 20% of democrats) voting on E day. There’s a logical argument to be made that the difference will be less pronounced than 2020. There’s a logical argument to be made that a bunch of first time voters will stumble into the polls on E day. But there’s no indication that the vote method split has completely subsided. If it’s 42-38 in Florida the entire time, the entire Republican ticket will win by literally double digits.

[Unelectable Bystander]

It’s not me saying that Dems will vote by mail, it’s the polls of Georgia/Texas/I think some other states where they actually said it. Coupling this with primary results, and the divide being present in primaries (for those  of you that say you are voting in person, I’m guessing you also did that in the primaries so it would be reflected in this analysis) AND special elections AND 2021 statewide elections, I think there’s a pretty good case that democrats will still dominate the early voting and get shellacked on E day. In fairness, Florida could probably somewhat of an exception with the GOP still being the party of high propensity voters there.

With that being said, a few days worth of data on 4 states is not conclusive and does not point to a definite GOP victory. It is, however, either a nothing burger or a good sign for the GOP to some degree. The fact that the first reaction of atlas is to declare that they are 100% sure it’s irrelevant shows that: A) if there is a wave, nobody will see it coming and B) this is a forum of advocacy more so than “looking at the data”

[Unelectable Bystander]

Georgia is looking grim.
https://twitter.com/winwithjmc/status/1583180701217984512?s=21&t=aqKKZz5SipC...

The same exact drop happened in 2018 and 2020 if I read the comments and the graphs posted correctly. In no sense is this “grim.”

What did this end up at? I feel like I keep seeing numbers with no reference of comparison.

That target smart graphic says that the early vote ended up right around 30% in 2020. They somehow have it at 37% right now though.

[Unelectable Bystander]

Georgia:

2022 Day 5 total EV compared to Day 5 total EV in:

2018, % of Electorate
+3 Black
-7 White*
+1 Female
-2 65 & up

2020, % of Electorate
+2 Black
-1 White
+2 Male
+17 65 & up

My estimates suggest the total EV at this point is around 4 points more Democratic than 2018 and 1 point more Republican than 2020 (margin) when taking an average of support levels for these four groups in 2018 & 2020.

And just for reference, the final EV figures (ABM + AIP) for the past three elections:

2021 Warnock +13.76
2020 Biden    +5.73
2018 Abrams +1.61

*I would take the white shift with a grain of salt, given the ever-increase of voters being labeled as "other/unknown" when first registering as opposed to their actual race or ethnicity. I imagine the real number when assigning everybody properly is probably 4-5 points less white than 2018 rather than 7-8

Are those early vote totals official? A recent Quinnipiac poll had Warnock winning like 66% of the mail-in vote, 59% of the early in person, and Walker winning 60% of the E-day vote. That would mean the total early vote is about Warnock +26 or double his 2021 margin. That doesn’t align with those historical results at all. If those results are official and the poll is correct, that might mean that the voting method divide is only growing but R’s are over-performing by staying near 2018 levels of early vote. Alternatively, the poll might be awful.

[Unelectable Bystander]

Does anybody have numbers from NJ and VA compared to both 2021 and 2020? They may not generalize to all states but could be helpful for two things:

A) Both states in 2021 were wave electorates (and 2 different turnout scenarios at that) in the GOP early voting paranoia era. It’d be interesting to see who is meeting/exceeding their benchmarks from 2021.

B) If comparing 2021 to 2020, was the early vote obvious that it would be a shellacking?

[Unelectable Bystander]

Checked the NC breakdown for the first time in a while and the Dems are still outperforming 2020 percentage wise.

Where is this coming from? I see republicans over-performing by both modeled and registered party per target smart.

[Unelectable Bystander]

These are my ratings at this moment. Ratings are not based on how the early vote or total vote will end up, but who it is favoring currently and how strong the evidence is.

Safe D: Michigan

Likely D: Georgia

Lean D: Wisconsin, Ohio, Pennsylvania, Texas

Tilt R: Nevada

Lean R: Iowa, Arizona

Likely R: North Carolina, a bunch of blue states

Safe R: Florida

[Unelectable Bystander]

https://twitter.com/SimonWDC/status/1584953574504960000

Basically red states and swing states are the bright spot for Dems currently. Blue states plus IA/NV/NC/AZ are the only places looking good for the GOP. I do think I recall that North Carolina saw a similar pattern at the beginning of their early vote where Dems were outpacing 2020 for a while, but the GOP has been steadily gaining (not just gaining percentage, which is to be expected, but gaining relative to the same time in 2020).

This should tell a lot in the next couple of weeks. If these numbers hold up and Dems are straight up winning the early vote in Texas or Ohio, a wave is probably out of the question. If those numbers become more R favorable as time goes on the same way NC has, it could be a sign that special elections/first few days of early voting metrics are being skewed by dem super voter intensity.

[Unelectable Bystander]

Not sure how NV and AZ necessarily look good for GOP. Jury is still largely out, but Dems still have a lead in NV with not that much mail in, and AZ's return rates are not very good for the GOP.

AZ because the graphic above is showing Dems leading about 41-36 when it was like 46-33 at this point in 2020. When you factor in the modeled preference of the independents, target smart actually has GOP leading currently. They mention Arizona being a rough spot in their tweet if you don’t believe me. NV is hard to gauge with such little data but the fact that Washoe is outvoting Clark currently helps the GOP. Both are obviously subject to change.

[Unelectable Bystander]

Not sure how NV and AZ necessarily look good for GOP. Jury is still largely out, but Dems still have a lead in NV with not that much mail in, and AZ's return rates are not very good for the GOP.

AZ because the graphic above is showing Dems leading about 41-36 when it was like 46-33 at this point in 2020. When you factor in the modeled preference of the independents, target smart actually has GOP leading currently. They mention Arizona being a rough spot in their tweet if you don’t believe me. NV is hard to gauge with such little data but the fact that Washoe is outvoting Clark currently helps the GOP. Both are obviously subject to change.

I see that, but TargetSmart you're also comparing 309k ballots to 860k ballots, so a more apt comparison would probably be when we're closer to that figure.

At the same time, 2018 shows 457K ballots with an R+18 lead at this time... and Sinema won that year obv.

I would advise against comparing anything to 2018. That was a time when early voting skewed old even more than it does now, and before Trump made some republicans paranoid about early voting. Some might think that 2020 is a valuable comparison and some might not. I personally do because I expect the partisan patterns to stay relatively similar,  based on 2020, primaries, and polling data (though I do understand the perspective of being skeptical of it).

Edit: I meant based on 2021

[Unelectable Bystander]

https://twitter.com/SimonWDC/status/1584953574504960000

Basically red states and swing states are the bright spot for Dems currently. Blue states plus IA/NV/NC/AZ are the only places looking good for the GOP. I do think I recall that North Carolina saw a similar pattern at the beginning of their early vote where Dems were outpacing 2020 for a while, but the GOP has been steadily gaining (not just gaining percentage, which is to be expected, but gaining relative to the same time in 2020).

This should tell a lot in the next couple of weeks. If these numbers hold up and Dems are straight up winning the early vote in Texas or Ohio, a wave is probably out of the question. If those numbers become more R favorable as time goes on the same way NC has, it could be a sign that special elections/first few days of early voting metrics are being skewed by dem super voter intensity.

I don't know why people trust Target Smart in states without party registration like TX. Their modeled party "data" is often horribly inaccurate.

Just using the reliable data we see evidence of a massive red wave. But using a model is like trusting a Monmouth poll.

Could certainly be true. Speaking of, they had a new update today. Some items of note:

R’s gained nationally, looking closer to 2020 by modeled partisanship

Rural vote is up by a few points nationally. TX, NC, VA are looking notably rural (this includes TX 15, 28, and 34 at the district level, subject to small-ish data)

Speaking of Texas, the dem modeled partisanship lead was a mirage. It’s now close to 2020 levels, about 48-41 R

AZ, NV (subject to mail-in ballot question marks)  FL, NC are still R bright spots. Rust belt and GA are still D bright spots. Wisconsin is notably less rural, though it is more suburban.

Confessional R’s that look notably strong: All of California, RGV, IA-3, FL-27, ME-2

[Unelectable Bystander]

Some context on the Georgia EV:

https://twitter.com/stphnfwlr/status/1585275202719694853

This is a very dumb point by him. Sure, make the point that it might just be republicans voting earlier in the cycle, but the point about county size is meaningless. Obviously Greene county has less people than Atlanta, that’s already baked into the equation.

[Unelectable Bystander]

How much of Nevada is up to date? Are we waiting on mail dumps, non-reporting rural counties, etc? Target smart shows Clark ahead of other counties at about 30% of total 2020 early vote, but the Clark firewall is only 5%.

[Unelectable Bystander]

People need to stop grouping AZ with bad for Ds list. That is simply not the case. They’re exceeding/meeting 2018 benchmarks thus far.

The confounding factor is GOP EDay turnout. It may seem surprising, but only around 17 percent of the GOP primary vote in Maricopa, for example, was banked on Election Day this year (source: https://elections.maricopa.gov/asset/jcr:03a2698d-ff49-49eb-a2b2-5b4f5fe608c8/08-02-2022-1%20Final%20Official%20Summary%20Report%20AUG2022.pdf). Over 80 percent of partisan republicans voted early despite Trump!

And 15 percent voted on EDay for 2018 GOP primary… so it’s not some huge jump (https://elections.maricopa.gov/asset/jcr:d0cf70c9-ee46-4d2f-a4f4-a789870d0ee8/08-28-2018%20Final%20Summary%20Report.pdf).

It seems safe to say Ds will barely be a blip on EDay, however.

Maricopa republicans skew old so their votes skew early. Looking statewide tells a different story. These are the % of vote share received on election day, based on my calculations from the state website:

Biden 2020: 7%
Trump 2020: 15%
Kelly 2022: 7%
Senate republicans 2022: 22%

I’m even seeing Masters and Lake specifically getting about 26% of their votes on election day. 2018 is not a valid comparison as the vote method divide continues to grow.

[Unelectable Bystander]

People need to stop grouping AZ with bad for Ds list. That is simply not the case. They’re exceeding/meeting 2018 benchmarks thus far.

The confounding factor is GOP EDay turnout. It may seem surprising, but only around 17 percent of the GOP primary vote in Maricopa, for example, was banked on Election Day this year (source: https://elections.maricopa.gov/asset/jcr:03a2698d-ff49-49eb-a2b2-5b4f5fe608c8/08-02-2022-1%20Final%20Official%20Summary%20Report%20AUG2022.pdf). Over 80 percent of partisan republicans voted early despite Trump!

And 15 percent voted on EDay for 2018 GOP primary… so it’s not some huge jump (https://elections.maricopa.gov/asset/jcr:d0cf70c9-ee46-4d2f-a4f4-a789870d0ee8/08-28-2018%20Final%20Summary%20Report.pdf).

It seems safe to say Ds will barely be a blip on EDay, however.

Maricopa republicans skew old so their votes skew early. Looking statewide tells a different story. These are the % of vote share received on election day, based on my calculations from the state website:

Biden 2020: 7%
Trump 2020: 15%
Kelly 2022: 7%
Senate republicans 2022: 22%

I’m even seeing Masters and Lake specifically getting about 26% of their votes on election day. 2018 is not a valid comparison as the vote method divide continues to grow.

Still, Maricopa is fairly instructive since it’s where the majority lives. I would be  surprised if Masters and Lake’s ED vote is 26 percent of their total when the ED vote was around 20 percent of the total primary vote — and you would expect those folks, primary participants, to be the diehards. But who knows

That doesn’t make any sense. Why would we look at one county when the entire state’s numbers are available? As for the reason why those two have higher %’s than the rest of the republicans, it’s because being farther right and being less urban now make one more likely to vote on Election Day. We’ve seen it in GOP primaries, general elections, and even democrat primaries since 2020. The assumption that Election Day turnout will look similar to 2018 is a bad assumption and it’s the reason why Arizona was called early before Trump made it so close by winning E day 2 to 1.

Even if those numbers don’t hold up and Masters gets the same 15% as Trump instead of 22% or 26%, he will still win E day by 2 to 1 margins, which is why Kelly needs Biden margins in the early vote.

[Unelectable Bystander]

There are two things right now that could use some further explanation:

A) What’s causing the difference between the modeled early vote and the peripheral numbers? Target smart has the modeled vote overall similar to 2020, except every Republican demographic is up by at least 2-3%. Whites, rural voters, men, married voters, older voters, and even registered republicans are all significantly increasing their share of the early vote. In Georgia for example, the top 10 turnout counties are all rural, yet the modeled vote is very democrat friendly. The map above is a good graphic showing this. What gives?

I’d like to say that the models are bad, but if that’s the case then it was probably equally bad in 2020. There has to be a different explanation. Are democrats with republican-leaning demographics disproportionately voting, and vice versa?

B) there’s a lot of discussion on RV vs LV. If there is a democratic turnout boost from Dobbs, would we be seeing it now or will they show up on Election Day? Was there precedence for an Election Day surge in the specials where Dems performed well?

Has anybody seen anything on these topics? If not, would there be interest in researching them? Recommend if so.

[Unelectable Bystander]

https://twitter.com/FLVoiceNews/status/1586029783582142464
FL Republicans keep adding

The others are right 2016, you’ve already shown that Florida is gone. The only interesting part is Miami-Dade and it’s implications on FL-27. At this point I think it flips.

[Unelectable Bystander]

I’m updating my early voting ratings. Reminder this is my opinion on how good the news is for each party at this time. Obviously not an actual contest rating and is subject to change.

Likely D: Michigan

Lean D: Georgia (D’s still winning the “model”, but rural turnout is still high. I might shift this if white turnout keeps rising), Iowa (D’s have caught up to 2020 and have been gaining)

Toss-up: Wisconsin (I didn’t realize early in-person had only just started, so probably too early), Pennsylvania (tracking with 2020), and Ohio (D’s are winning the model, but urban and minority turnout are way down)

Lean R: Arizona, Texas (rural turnout is very high), Nevada (R’s are outpacing 2020 despite some rural counties reporting nothing)

Likely R: North Carolina (same story as Texas, but more data and registered R’s are outpacing 2020), Oregon, California

Safe R: Florida

[Unelectable Bystander]

I’m updating my early voting ratings. Reminder this is my opinion on how good the news is for each party at this time. Obviously not an actual contest rating and is subject to change.

Likely D: Michigan

Lean D: Georgia (D’s still winning the “model”, but rural turnout is still high. I might shift this if white turnout keeps rising), Iowa (D’s have caught up to 2020 and have been gaining)

Toss-up: Wisconsin (I didn’t realize early in-person had only just started, so probably too early), Pennsylvania (tracking with 2020), and Ohio (D’s are winning the model, but urban and minority turnout are way down)

Lean R: Arizona, Texas (rural turnout is very high), Nevada (R’s are outpacing 2020 despite some rural counties reporting nothing)

Likely R: North Carolina (same story as Texas, but more data and registered R’s are outpacing 2020), Oregon, California

Safe R: Florida

And these rankings are purely based on EV data?

Yes, in person and mail from target smart

[Unelectable Bystander]

@Unelectable Bystander,
I think Ohio & North Carolina is locked in for Republicans too.

The current D-R Spread in Ohio is 45-41 and in North Carolina if you believe the NC State Board of Elections it's 39-31 D-R down from 43-27 D-R in 2020.

The LAZY North Carolina Republican Party trends to get their Early Voters more out during the Final Week of EV.

I am pretty comfortable that those Spreads in OH & NC are not good enough to safe Tim Ryan and Cheri Beasley.

You keep saying this; it's not true. The final EV (in person + mail) was D+5, 37-32 in 2020

Both are right, it’s just different metrics. They’re not quite at what D’s finished 2020 with, but they are on pace if the pattern holds.

[Unelectable Bystander]

@Unelectable Bystander,
I think Ohio & North Carolina is locked in for Republicans too.

The current D-R Spread in Ohio is 45-41 and in North Carolina if you believe the NC State Board of Elections it's 39-31 D-R down from 43-27 D-R in 2020.

The LAZY North Carolina Republican Party trends to get their Early Voters more out during the Final Week of EV.

I am pretty comfortable that those Spreads in OH & NC are not good enough to safe Tim Ryan and Cheri Beasley.

You keep saying this; it's not true. The final EV (in person + mail) was D+5, 37-32 in 2020

Both are right, it’s just different metrics. They’re not quite at what D’s finished 2020 with, but they are on pace if the pattern holds.

Huh? There's only one factual # here - it's the D+5 final. Where is the 43-27 coming from?

That was their lead with 11 days to go. It implies that republicans will gain between then and Election Day. Not a guarantee, but that was the 2020 pattern

[Unelectable Bystander]

There are two things right now that could use some further explanation:

A) What’s causing the difference between the modeled early vote and the peripheral numbers? Target smart has the modeled vote overall similar to 2020, except every Republican demographic is up by at least 2-3%. Whites, rural voters, men, married voters, older voters, and even registered republicans are all significantly increasing their share of the early vote. In Georgia for example, the top 10 turnout counties are all rural, yet the modeled vote is very democrat friendly. The map above is a good graphic showing this. What gives?

I’d like to say that the models are bad, but if that’s the case then it was probably equally bad in 2020. There has to be a different explanation. Are democrats with republican-leaning demographics disproportionately voting, and vice versa?

B) there’s a lot of discussion on RV vs LV. If there is a democratic turnout boost from Dobbs, would we be seeing it now or will they show up on Election Day? Was there precedence for an Election Day surge in the specials where Dems performed well?

Has anybody seen anything on these topics? If not, would there be interest in researching them? Recommend if so.

I did some preliminary research into these.

A) I still don’t see a clear answer on this. One idea was to look at it by county vs overall turnout, to see if there’s anything obvious. For example, when we say “suburban”, are the more D-leaning counties turning out more than the R-leaning ones? So far I’ve looked at Ohio. Relative to the state:
3 C’s: Cuyohoga is near the state average, the other two are lagging far behind (this was surprising to me)
Trumbull and Lorain: Monster turnout so far. I think one of them is even approaching 2020 pace, while the state average is about 30% of 2020 EV.
Lucas: Around average
Warren/Butler: Very strong
Delaware: Around average

Nothing there answers the question for me on why the model has R’s underperforming. Hopefully more to come on other states.

B) Unfortunately, I don’t think that Kansas has results by vote method. This would have been very informative.

I was able to find results for Nebraska. They do not have vote method by party, but they have it broken out by county. Remember that most of the counties actually swung right here. The big difference was that Lancaster shifted slightly left while accounting for more vote share. My idea was to compare Lancaster to the random rurals which swung right. Did Lancaster have a huge jump in election day votes compared to the others?

The average rural county voted maybe 50-60% E-day in 2020, jumped to around 80% in the 2022 primary, then dropped slightly to around 75% for the special.

Lancaster was about 45% E-day in 2020, 50% in the primary, and 55% for the special. So there was a jump from the primary to the special in Lancaster compared to the others, but not really a disproportional jump from 2020 to the special.

Based on this, I don’t think I can prove or disprove a “hidden” Election Day Dem vote after Dobbs. I’m hoping to look at New York eventually also.