Early & Absentee Voting Megathread - Build the Freiwal

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

Is there a universe in which Latino turnout surges but they don't lean heavily D?

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

And in addition to the favorable party registration of early voters in NV, Rs are cannibalizing more to get their increased early vote, whereas Dems are bringing out more new/unlikely voters. It doesn't take a genius to see who is going to win this one.

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

Here is a breakdown of party registration of people who early voted in the NYT/Siena Arizona Senate poll. I posted this in one of the AZ poll threads, but am cross-posting here for its relevance to early voting as well.

It suggests that people with no party affiliation are breaking to Sinema.

I know everyone says “early voting numbers don’t matter” and the at might be accurate in other states but in AZ early voting is 80% of the election and as much of a proud dem in AZ that I am, we aren’t doing so good.

AZ-08 AZ-08 AZ-08

how many times does this need to be said

Wrong, CNN said Sinema lead Early Vote 54-43 and Marist said 51-47. Both of these Polls are just pure FANTASY.

GOP currently has a 116K Ballot Advantage in AZ. For starters for Sinema to be tied in the Early Vote she would need to have NPA Voters 70-30 in her favour. That is not happening, no way. ZERO Chance. Even if Sinema holds D's in Early Voting 94-6 and McSally holds Republicans let's say 88-12 Sinema still would need NPA Voters to break her way 68-32 or something like that.

So, the notion that Sinema is leading EV is completely Baloney.

FWIW, in the NYT/Siena poll, if you look at the microdata you can see how people with different party registrations who said they early voted voted, to see how they came up with McSally leading early voters. Caveat being that this is a small sample size, but if you want to actually see how they got their #s, you can see.

Counting unweighted #s of people, out of 178 people in the poll who said they early voted, 69 (39%) were registered Rs, 58 (33%) were registered Ds, and 51 (29%) were registered Ds.

The registered Ds split their vote 49-7-2 (Sinema-McSally-Refused_to_say)
The registered Rs split their vote 8-59-2 (Sinema-McSally-Refused_to_say)
The registered Indepents/NPAs split their vote 32-18-1 (Sinema-McSally-Refused_to_say)

In percentage terms, that is Sinema winning registered Ds 84%-12%, McSally winning registered Rs 86%-12%, and Sinema winning registered Independents/NPAs 63%-35%.

And overall, that comes out to Sinema winning the early voters 50%-47%.

Then if you apply the NYT/Siena polls weighting, that gets you to Sinema leading among early voters 51%-45%.

Granted, this does not prove that Sinema is leading with early voters, and these are small sub-samples.  But this is how Sinema can be leading with early voters despite Rs having a registration advantage. And indeed, Sinema is handily winning voters with no party affiliation in the sample, which is how she can win even if there is an R registration advantage. Granted also the sample could be somewhat off, and it is quite possible that McSally could be winning early voters (but probably not by much). We have had several polls all with Sinema winning or competitive among the early vote subsample. That does suggest that despite party registration, Sinema will at least be fairly competitive in the early vote, and then we have to see what the election day vote is like.

You are just being deliberately obtuse or willfully blind, you are smart enough that there is no way that you do not understand this.

Party registration /= votes. Duh.

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

That sample is way off by party ID though, particularly in terms of overrepresenting NPAs. In actuality the early vote % is 42.3-33.7-23.3 right now, no where close to 39-33-29. If you assume those %s on the actual party ID of early voters you get McSally 48.6 - Sinema 48.0%. And that’s assuming Sinema is winning NPAs by 28 points.

I’m not saying it is impossible for Sinema to win, but in order to win the early vote she has to win NPAs by over thirty or do substantially better among Ds than McSally does with Rs. And of course getting a boost with the final few days of early voting would help too.

I don't disagree that NPAs are slightly overrepresented in the sample (by a few points), but I wouldn't call that *way* off given that we are talking about a small sub-sample. Remember also that deviations in party registration of early voters tend to get counteracted by weighting of people in the rest of the poll who are not early voters.

And I also am not saying that Sinema will definitely win early voters, just that the polling suggests she will be at least reasonably competitive among them despite the R party registration advantage. The overriding point is simply that the R party registration advantage among early voters so far does not mean that Sinema is not favored, much less that she can't win. Of course McSally could win too, nobody (at least nobody sensible) is arguing that that is not possible. But some Rs are getting pretty desperate to unskew every AZ poll they can see, on the grounds that supposedly McSally can't lose if there is an R registration advantage.

Like I said:

Granted also the sample could be somewhat off, and it is quite possible that McSally could be winning early voters (but probably not by much). We have had several polls all with Sinema winning or competitive among the early vote subsample. That does suggest that despite party registration, Sinema will at least be fairly competitive in the early vote, and then we have to see what the election day vote is like.

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

https://twitter.com/RalstonReports/status/1058430087844945920

Looks like the freiwal's gonna get 10 feet taller today.

You ask what height will the Freiwal be? It will be Frei Heit!

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

LAWD JESUS

https://mobile.twitter.com/zachdespart/status/1058532621192515586

For some context on that, 105,005 people voted on the last day of early voting in Harris County in 2016. So if they still have long lines, they will probably pretty much meet and perhaps exceed that # today.

And that 105,005 number is, I think, the all time record for highest turnout in a day of early voting in a TX county.

51,628 voted on the last day of early voting in 2014.

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

Looks great for Rosen.

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

So does anyone know off hand what the # was for the last day of early voting in Clark County in 2016? Any chance of beating that now?

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

Massive line in Henderson:

https://twitter.com/k_newberg/status/1058552546027962368

Silverado Ranch is a GOP heavy area I think so may not be that good of news.

It is in a D+12.45 precinct on dave's redistricting App. The precinct it is in borders precincts which are D+4.99, D+8.9, D+11.6, D+5, and there is one R+10.07 precinct. There are some more R+ precincts in the general area (particularly to the South), but there are plenty of D+ as well, and the D+ precincts tend to be nearer by. At the very worst, it is competitive.

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

Massive line in Henderson:

https://twitter.com/k_newberg/status/1058552546027962368

Silverado Ranch is a GOP heavy area I think so may not be that good of news.

Someone posted a precinct map and said it was 60-34 Hillary.

Is this the place they're talking about?  It looks like it's in Enterprise, not Henderson.

Yeah, that is the same one I was looking at in DRA, I am pretty sure it is that one.

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

https://mobile.twitter.com/mattlargey/status/1058568111488724992

2016 last day was 42,932
2014 last day was 22,059

So pretty good - very close to 2016 turnout there.

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

Getting hard to see how anyone can call NV a tossup at this any point

It is hard to see how anyone can call NJ or TX a tossup too, but that doesn't stop the Cook Political Report.

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

I made some maps comparing the 2018 Clark County EV+MB vote to the 2016 EV+MB:

Nice maps, well done!

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

It's really odd to me why people continue to act like Beto can't win. Sure, is it likely? Probably not. But it's not impossible. I don't get why people here are instantly writing it off. Bad analyzation.

The argument for Cruz winning: Numbers.

The argument for Beto winning: Hopey-changey stuff.

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

Still won't make a difference in the end. Cruz will win.

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

Final update on TargetSmart nationwide early vote data (Focusing on the crosstab of Partisanship and Vote History)


Here is a final look at the early vote data made available by TargetSmart. To me, the thing particularly important to focus on in looking at data is partisanship and vote likelihood, and comparing how this has changed across different vote history groups and across different states. The data is either complete or almost complete in every state now. So let's just jump right in for one last look at what early vote can tell us!

And as a reminder, this data is modeled partisanship in the voter file - it is meant to predict whether someone generally thinks of themselves as a D or an R, but not per se which individual candidates they will necessarily vote for. Some candidates will obviously get better %s of the vote than others, and polling is the best guide to that.

https://targetearly.targetsmart.com/index.html

The first set of numbers below ("45.5% D --- 46.2% R --- 8.3% UNK") means that across the entire USA, in all the early votes TargetSmart has collected, 45.5% of voters have modeled Dem partisanship, 46.2% have modeled Republican partisanship, and 8.3% have unknown modeled partisanship. The second set of numbers ("D+3.8 Swing in Partisanship Margin from 2014") is the size of the swing in early voter turnout as compared to 2014. So, for example, among all voters, in the 2018 data modeled Reps outnumbered modeled Dems by .7%. But in 2014, the numbers were 48.4% Modeled R to 43.9% Modeled D, a Republican advantage of 4.5%. So the swing from a Modeled Republican advantage of 4.5% in 2014 to a Modeled Republican advantage is 3.8%, which is what the "D+3.8 Swing in Partisanship Margin from 2014" means.

So, on its face, it may seem good for Republicans that they have a slight advantage in Modeled partisanship of early voters. However, it is good for Dems that there is a significant swing in favor the Democrats, which indicates better Dem turnout in 2018 early voting than in 2014, even though Rep turnout has also improved, Dems have improved by more overall.

Code:

USA --- All Voters  ---           45.5% D --- 46.2% R --- 8.3% UNK ---                          D+3.8 Swing in Partisanship Margin from 2014

The better news for Democrats, however, comes when you break the overall early vote turnout down by likelihood of voting. Simply put, Democrats have improved by the most, and are doing the best, in turning out less likely voters, while Republicans are doing (relatively) better in turning out their consistent, reliable voters.

Code:

USA --- Super Voter ---           43.6% D --- 50.9% R --- 5.5% UNK ---                          D+1.6 Swing in Partisanship Margin from 2014
USA --- Freq. Voter ---           45.6% D --- 46.2% R --- 8.3% UNK ---                          D+2.8 Swing in Partisanship Margin from 2014
USA --- InFreq. Vt. ---           48.8% D --- 39.0% R --- 12.1% UNK ---                          D+5.2 Swing in Partisanship Margin from 2014
USA --- Never Voted ---           47.3% D --- 36.3% R --- 16.4% UNK ---                          D+4.8 Swing in Partisanship Margin from 2014

Republicans have a 50.9% to 43.6% advantage among Super Voters (approximately the top quintile of voters in terms of their likelihood to vote), but that is relatively less meaningful than the Dem advantage among less likely voters, because everyone knew all along that the great majority of Super Voters are going to vote - it was just a question of whether they vote early or on election. Republicans do also have a narrow advantage (46.2% to 45.6%) among Frequent voters. But Dems have bigger advantages (and have improved the most relative to 2014) among Infrequent voters (48.8% to 39.0% Modeled Dem lead, and a swing of D+5.2%) and people who have never voted before in any election (47.3% to 36.3% Modeled Dem lead, and a swing of D+4.8%).

The optimistic way for Republicans to interpret this is to hope that the swings are not "too big." The optimistic way for the Dems to interpret this is that there are significant swings in turnout in Dems favor, and in particular that Republicans are in relatively greater danger of having cannibalized their election day vote than Dems are, because Republicans are more reliant on Super Voters (and Frequent Voters) than Dems, and Dems have been doing a better job of turning out Infrequent Voters and New Voters.


Now let's look at individual states... Arizona first.

Code:

AZ --- All Voters  ---           40.8% D --- 53.5% R --- 5.7% UNK ---                          D+5.1 Swing in Partisanship Margin from 2014
AZ --- Super Voter ---           35.8% D --- 61.7% R --- 2.5% UNK ---                          R+2.7 Swing in Partisanship Margin from 2014 ---              (D+1.1 Change in Swing Since Oct 31)
AZ --- Freq. Voter ---           40.5% D --- 53.5% R --- 6.0% UNK ---                          D+2.4 Swing in Partisanship Margin from 2014 ---              (D+0.9 Change in Swing Since Oct 31)
AZ --- InFreq. Vt. ---           46.7% D --- 44.2% R --- 9.0% UNK ---                          D+13.3 Swing in Partisanship Margin from 2014 ---              (D+1.4 Change in Swing Since Oct 31)
AZ --- Never Voted ---           51.5% D --- 38.0% R --- 10.5% UNK ---                          D+23.5 Swing in Partisanship Margin from 2014 ---              (R+1.1 Change in Swing Since Oct 31)

Overall, Rs have a solid 53.5% - 40.8% lead in modeled partisanship of early voters. However, Republicans typically do well among early voters, and this represents a 5.1% improvement for Modeled Dems from 2014 (1.3% larger than the national average swing).

Moreover, the Republican advantage is particularly strong among Super voters. Despite Dems improving by more than the average overall, there was actually a swing to Republicans of R+2.7 among Super voters. Republican likely voters in AZ have early voted like crazy. Which on the one hand is good for Republicans, since they did actually vote... But it means that Republicans are particularly vulnerable to cannibalization of the election day vote. Meanwhile, Dems had HUGE improvements (D+13.3 and D+23.5) in turning out Infrequent Voters and people who never voted before. This mean that when you consider the vote history of people who have early voted, the early voting in AZ looks quite a bit better than if you only look at the overall topline R advantage among all voters.

So, whatever the initial early vote returns are in AZ, there is very good reason to think that, unless there is really low election day turnout and additional Dems barely vote on election day, Republicans will do relatively well in the early vote, and then Dems will improve as more of the election day vote comes in. High election day turnout = good for Dems, low election day turnout = good for Reps.

The last set of numbers ("D+1.1 Change in Swing Since Oct 31") is a comparison to the last time I looked at the TargetSmart Data on 2014. So for example, Dems have improved (relative to 2014) by 1.1%. HOWEVER, the previous numbers represent how things were going on the SAME DAY in 2014, not on the CURRENT DAY in 2014, so this is really a measurement of whether Dems have improved BY MORE THAN THEY IMPROVED OVER THE LAST FEW DAYS IN 2014, as opposed to a measurement of whether Dems have improved in absolute terms. So it is possible for these numbers to be R+, but for Dems to have still improved in absolute terms, if they improved by less than they improved over the last few days in 2014.

Anyway, so in the case of AZ, we can see that these last set of numbers comparing to Oct 31 generally are D+, indicating that Dems have been improving in late-returned ballots by a bit more than was the case in 2014.

Code:

CA --- All Voters  ---           59.1% D --- 36.3% R --- 4.7% UNK ---                          D+10 Swing in Partisanship Margin from 2014
CA --- Super Voter ---           54.3% D --- 43.1% R --- 2.6% UNK ---                          D+8.8 Swing in Partisanship Margin from 2014 ---              (D+2.5 Change in Swing Since Oct 31)
CA --- Freq. Voter ---           59.8% D --- 35.3% R --- 5% UNK ---                          D+9.5 Swing in Partisanship Margin from 2014 ---              (D+3.8 Change in Swing Since Oct 31)
CA --- InFreq. Vt. ---           65.8% D --- 26.9% R --- 7.3% UNK ---                          D+6.6 Swing in Partisanship Margin from 2014 ---              (D+3 Change in Swing Since Oct 31)
CA --- Never Voted ---           66.9% D --- 23.4% R --- 9.8% UNK ---                          D+2.3 Swing in Partisanship Margin from 2014 ---              (D+4.8 Change in Swing Since Oct 31)

In CA, Dems are cleaning up as we would expect. They are, however, at a bit more risk of cannibalization than in 2014, because Dems have improved more among Super Voters and Frequent Voters.

Of course, the most important question is really what is happening within the competitive Congressional districts, and this doesn't tell us that much about that, since these are statewide numbers.

Dems are also improving more in the late returned ballots since Oct 31 than was the case in 2014.

Code:

CO --- All Voters  ---           46.5% D --- 45.9% R --- 7.6% UNK ---                          D+8.1 Swing in Partisanship Margin from 2014
CO --- Super Voter ---           44.0% D --- 51.9% R --- 4.1% UNK ---                          D+5.4 Swing in Partisanship Margin from 2014 ---              (R+1.8 Change in Swing Since Oct 31)
CO --- Freq. Voter ---           47.0% D --- 44.2% R --- 8.7% UNK ---                          D+7.9 Swing in Partisanship Margin from 2014 ---              (R+0.7 Change in Swing Since Oct 31)
CO --- InFreq. Vt. ---           49.9% D --- 38.9% R --- 11.1% UNK ---                          D+13.2 Swing in Partisanship Margin from 2014 ---              (D+1.2 Change in Swing Since Oct 31)
CO --- Never Voted ---           52.4% D --- 33.9% R --- 13.7% UNK ---                          D+12 Swing in Partisanship Margin from 2014 ---              (D+1.1 Change in Swing Since Oct 31)

Dems look very solid in CO. Big improvements among all types of voters, but biggest among Infrequent and New voters, so not much cannibalization risk either. CO Dems seem set to do very well.

Code:

FL --- All Voters  ---           46.0% D --- 48.3% R --- 5.7% UNK ---                          D+5.3 Swing in Partisanship Margin from 2014
FL --- Super Voter ---           43.5% D --- 53.4% R --- 3.1% UNK ---                          D+0.8 Swing in Partisanship Margin from 2014 ---              (R+4.2 Change in Swing Since Oct 31)
FL --- Freq. Voter ---           45.7% D --- 48.1% R --- 6.2% UNK ---                          D+6 Swing in Partisanship Margin from 2014 ---              (R+3.1 Change in Swing Since Oct 31)
FL --- InFreq. Vt. ---           50.5% D --- 41.2% R --- 8.3% UNK ---                          D+7.9 Swing in Partisanship Margin from 2014 ---              (R+1.8 Change in Swing Since Oct 31)
FL --- Never Voted ---           50.0% D --- 39.1% R --- 11% UNK ---                          D+14.9 Swing in Partisanship Margin from 2014 ---              (R+0.3 Change in Swing Since Oct 31)

The early vote vote in Florida seems to be looking pretty solid for Dems (small 48.3% to 46.0% R advantage, but with an overall D+5.3 swing). The especially good part is that the Dem improvement is strongly concentrated among Infrequent Voters, New Voters, and to some extent Frequent Voters.

So from this data, I would be pretty confident about Gillum/Nelson winning. In 2016, Trump famously won Florida by doing better than expected in the Election Day vote. That seems less of a risk this time because the Dem improvement is so strongly concentrated among Infrequent Voters and New Voters. Republicans are at more relative risk of cannibalization. So in Florida, high election day turnout should be good for Dems. Low election day turnout should be good for Reps.

However, there is one bit of bad news for Dems here, and that is that since Oct 31, the swings as compared to the same time in 2014 are all R+. That means Dems have improved less than in 2014 in the late early voting.

Code:

GA --- All Voters  ---           44.4% D --- 50.1% R --- 5.5% UNK ---                          D+4.2 Swing in Partisanship Margin from 2014
GA --- Super Voter ---           39.2% D --- 58.2% R --- 2.6% UNK ---                          D+3.4 Swing in Partisanship Margin from 2014 ---              (R+0.4 Change in Swing Since Oct 31)
GA --- Freq. Voter ---           46.7% D --- 48.6% R --- 4.7% UNK ---                          R+0.1 Swing in Partisanship Margin from 2014 ---              (D+0.1 Change in Swing Since Oct 31)
GA --- InFreq. Vt. ---           47.3% D --- 43.0% R --- 9.7% UNK ---                          R+4.2 Swing in Partisanship Margin from 2014 ---              (D+2.3 Change in Swing Since Oct 31)
GA --- Never Voted ---           46.7% D --- 39.5% R --- 13.8% UNK ---                          D+0.2 Swing in Partisanship Margin from 2014 ---              (D+1.7 Change in Swing Since Oct 31)

Georgia is one of the few states where looking closely at the early vote data makes me less optimistic about Dem chances than I otherwise would be.

There *is* a 4.2% overall swing in Partisanship of early voters as compared to 2014. However, unfortunately Georgia is very different from most other states in that the Dem swing has come basically entirely from Super Voters. This should be pretty surprising, since turning out low propensity voters has been the #1 focus of the Abrams campaign. But it looks like a lot of low propensity Republicans are also turning out. That means Dems are at risk of having cannibalized the election day vote in Georgia, and I am pretty worried about the election day vote. I suspect that, and given that African Americans who are eligible to vote were already very well mobilized compared to most other states, and given felon disenfranchisement laws that disproportionately effect African American males, there are just not enough additional low-propensity Dems to secure a victory by itself, unless low propensity White Republicans don't turn out.

One other way to spin things onto the bright side for Dems is that Dems have improved among Infrequent and New voters by D+2.3 and D+1.7 since Oct 31 (compared to the same time in 2014). The optimistic hope would be that this is a forerunner of huge election day turnout among Dem-leaning Infrequent and New voters. But overall, I would be pretty worried that Republicans seem to have done a pretty good job of turning out their Infrequent voters in Georgia, and that too much of the Dem early vote turnout may be cannibalization of Election Day super-voters, as opposed to actually bringing in new voters.

The only other optimistic take I can see for Dems is that maybe a lot of the additional Modeled R partisanship voters are not really Rs any more. If a lot of those are actually former Rs in the North Atlanta suburbs, but they swing en masse to Abrams, then things could go better than the data seems to indicate. If Abrams does win, I think it will be more because of a large swing among white voters in the North Atlanta suburbs towards Dems, as opposed to just Abrams turning out the base.

Regretfully, I am thinking that despite Abrams doing well in the polls, I may need to rate GA-GOV as tilt R for my final prediction, though I have not made a final decision yet. But this is really one state where it will all come down to election day turnout, and most likely also to a runoff.

Code:

IA --- All Voters  ---           46.8% D --- 44.2% R --- 9.0% UNK ---                          D+2.4 Swing in Partisanship Margin from 2014
IA --- Super Voter ---           47.5% D --- 47.3% R --- 5.2% UNK ---                          D+4.8 Swing in Partisanship Margin from 2014 ---              (R+0.8 Change in Swing Since Oct 31)
IA --- Freq. Voter ---           45.2% D --- 44.1% R --- 10.6% UNK ---                          R+1.4 Swing in Partisanship Margin from 2014 ---              (D+0.3 Change in Swing Since Oct 31)
IA --- InFreq. Vt. ---           49.2% D --- 38.5% R --- 12.3% UNK ---                          D+5.1 Swing in Partisanship Margin from 2014 ---              (D+0.5 Change in Swing Since Oct 31)
IA --- Never Voted ---           47.5% D --- 38.1% R --- 14.3% UNK ---                          D+1.1 Swing in Partisanship Margin from 2014 ---              (D+0.7 Change in Swing Since Oct 31)

Iowa looks OK for Dems. There is a swing to Dems from 2014, but less than the national average. Strangely, the Dem advantage is concentrated among Super Voters AND Infrequent Voters, whereas Rs are doing relatively better with Frequent voters AND New voters. So I have no idea what that means for potential cannibalization.

Code:

IL --- All Voters  ---           56.6% D --- 30.0% R --- 13.5% UNK ---                          D+6.8 Swing in Partisanship Margin from 2014
IL --- Super Voter ---           56.8% D --- 31.6% R --- 11.6% UNK ---                          D+4.5 Swing in Partisanship Margin from 2014 ---              (R+1.7 Change in Swing Since Oct 31)
IL --- Freq. Voter ---           55.0% D --- 31.1% R --- 13.9% UNK ---                          D+6.5 Swing in Partisanship Margin from 2014 ---              (R+2.2 Change in Swing Since Oct 31)
IL --- InFreq. Vt. ---           60.1% D --- 23.4% R --- 16.5% UNK ---                          D+14.7 Swing in Partisanship Margin from 2014 ---              (R+2.1 Change in Swing Since Oct 31)
IL --- Never Voted ---           57.0% D --- 22.9% R --- 20.1% UNK ---                          D+2 Swing in Partisanship Margin from 2014 ---              (D+0.4 Change in Swing Since Oct 31)

IL looks fabulous for Dems. There is a larger than average. We don't know exactly how this breaks down by Congressional District, but if I were an R Congressional incumbent in any one of IL-06, IL-12, IL-13, or IL-14 looking at these numbers, I would be VERY worried by them. If this is at all an indication of the final results, the Pritzkerslide could be big and bad for IL Republicans.

Code:

MI --- All Voters  ---           32.6% D --- 50.2% R --- 17.3% UNK ---                          R+2.2 Swing in Partisanship Margin from 2014
MI --- Super Voter ---           33.0% D --- 52.9% R --- 14.2% UNK ---                          D+0.1 Swing in Partisanship Margin from 2014 ---              (R+2.1 Change in Swing Since Oct 31)
MI --- Freq. Voter ---           31.6% D --- 49.0% R --- 19.4% UNK ---                          R+5.7 Swing in Partisanship Margin from 2014 ---              (R+2.1 Change in Swing Since Oct 31)
MI --- InFreq. Vt. ---           34.6% D --- 41.9% R --- 23.5% UNK ---                          R+7.9 Swing in Partisanship Margin from 2014 ---              (R+3.2 Change in Swing Since Oct 31)
MI --- Never Voted ---           29.8% D --- 38.3% R --- 31.9% UNK ---                          R+9.7 Swing in Partisanship Margin from 2014 ---              (R+4.2 Change in Swing Since Oct 31)

Given how good the polls are in MI for Dems, the early vote looks strangely good for Republicans. If there were a single state where one might predict Republicans overperforming expectations based on the election day vote, it would probably have to be Michigan.

This is one of the very few states where R partisanship among early voters has actually improved over 2014 (R+2.2). And what is even worse for Dems, the Republican gains are concentrated among Infrequent and New voters, so Dems may be cannibalizing their election day vote in Michigan more (as compared to 2014). And Republicans have also been improving in the late returns over the last few days by relatively more than 2014 (or Dems improving by relatively less than 2014).

The only good interpretation of this for Dems is that early vote is not a big thing in Michigan, so hopefully the early voting seeming relatively good for Republicans will be washed away by an overwhelming election day Whitmerslide. Nevertheless, this is 874k votes we are talking about, which is definitely not nothing.

Otherwise, these are not good numbers for Dems.

Anyway, so if there is a surprise where Republicans do better than expected in one particular state, I am eerily afraid it might be in Michigan again.

Code:

MN --- All Voters  ---           42.7% D --- 33.8% R --- 23.5% UNK ---                          D+0.3 Swing in Partisanship Margin from 2014
MN --- Super Voter ---           43.6% D --- 32.9% R --- 23.5% UNK ---                          R+2.4 Swing in Partisanship Margin from 2014 ---              (D+0.6 Change in Swing Since Oct 31)
MN --- Freq. Voter ---           38.5% D --- 37.4% R --- 24.1% UNK ---                          D+0.6 Swing in Partisanship Margin from 2014 ---              (D+2.1 Change in Swing Since Oct 31)
MN --- InFreq. Vt. ---           48.4% D --- 29.1% R --- 22.5% UNK ---                          D+6.9 Swing in Partisanship Margin from 2014 ---              (D+3 Change in Swing Since Oct 31)
MN --- Never Voted ---           52.5% D --- 23.9% R --- 23.6% UNK ---                          D+11 Swing in Partisanship Margin from 2014 ---              (D+1.3 Change in Swing Since Oct 31)

Minnesota has a smaller than average (D+0.3) overall improvement for Dems in the early vote. However, the Dem gains are concentrated among unlikely voters, and Republicans have actually improved (R+2.4) among Super Voters. So at least on the bright side Republicans are at a bit more risk of election day cannibalization of Dems.

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

Final update on TargetSmart nationwide early vote data (Focusing on the crosstab of Partisanship and Vote History) CONTINUED[size]

Code:

MT --- All Voters  ---           29.3% D --- 46.7% R --- 24.0% UNK ---                          D+1.9 Swing in Partisanship Margin from 2014
MT --- Super Voter ---           27.7% D --- 48.7% R --- 23.6% UNK ---                          R+3.1 Swing in Partisanship Margin from 2014 ---              (R+1.3 Change in Swing Since Oct 31)
MT --- Freq. Voter ---           26.2% D --- 50.3% R --- 23.4% UNK ---                          D+1.4 Swing in Partisanship Margin from 2014 ---              (R+0.6 Change in Swing Since Oct 31)
MT --- InFreq. Vt. ---           34.8% D --- 41.6% R --- 23.6% UNK ---                          D+5.2 Swing in Partisanship Margin from 2014 ---              (D+0.8 Change in Swing Since Oct 31)
MT --- Never Voted ---           41.3% D --- 25.9% R --- 32.8% UNK ---                          D+5.2 Swing in Partisanship Margin from 2014 ---              (D+2.8 Change in Swing Since Oct 31)

Like MN, Montana has smaller than average Dem gains as compared to 2014 (D+1.9). However, the Dem gains are a bit bigger overall than in MN. Rs have improved with Super Voters, Ds have improved with Infrequent and New voters - so there may be a bit of R cannibalization in the early vote. Another thing common to MN and MT, however, is that there are an above average # of people with Unknown partisanship. That is because neither state has Party Registration, and because most voters are white (so you can't tell by a voter's race whether they are likely to be D or R as in many Southern states that don't have Party registration).

Code:

NC --- All Voters  ---           51.6% D --- 43.4% R --- 5.0% UNK ---                          D+0.5 Swing in Partisanship Margin from 2014
NC --- Super Voter ---           52.1% D --- 44.6% R --- 3.3% UNK ---                          R+2.1 Swing in Partisanship Margin from 2014 ---              (R+0.9 Change in Swing Since Oct 31)
NC --- Freq. Voter ---           50.6% D --- 44.7% R --- 4.7% UNK ---                          D+1.2 Swing in Partisanship Margin from 2014 ---              (R+0.3 Change in Swing Since Oct 31)
NC --- InFreq. Vt. ---           52.7% D --- 39.3% R --- 8.1% UNK ---                          D+2.3 Swing in Partisanship Margin from 2014 ---              (D+0.4 Change in Swing Since Oct 31)
NC --- Never Voted ---           53.8% D --- 34.4% R --- 11.8% UNK ---                          D+8.9 Swing in Partisanship Margin from 2014 ---              (D+1.5 Change in Swing Since Oct 31)

There's not that big of a change from 2014 in NC. A bit in favor of Dems, and maybe a bit of R cannibalization. But NC is going to be a weird state, because there is no Senate race and no Governor's race. So really, the statewide #s don't tell us that much, and we should be more interested in how these numbers look like in the handful of Congressional Districts that are being seriously contested.

Code:

ND --- All Voters  ---           12.5% D --- 67.7% R --- 19.9% UNK ---                          R+4.8 Swing in Partisanship Margin from 2014
ND --- Super Voter ---           12.1% D --- 68.2% R --- 19.6% UNK ---                          R+8.7 Swing in Partisanship Margin from 2014 ---              (R+4.8 Change in Swing Since Oct 31)
ND --- Freq. Voter ---           11.7% D --- 68.7% R --- 19.6% UNK ---                          R+3 Swing in Partisanship Margin from 2014 ---              (R+0.5 Change in Swing Since Oct 31)
ND --- InFreq. Vt. ---           15.2% D --- 63.4% R --- 21.4% UNK ---                          D+1.9 Swing in Partisanship Margin from 2014 ---              (D+2.4 Change in Swing Since Oct 31)
ND --- Never Voted ---           32.6% D --- 39.9% R --- 27.5% UNK ---                          D+31.6 Swing in Partisanship Margin from 2014 ---              (D+18.8 Change in Swing Since Oct 31)

ND simply does not look good for Dems in this data. Rs may be cannibalizing a bit, but still... It is just not good looking for Heitkamp. Don't pay attention to the "D+31.6" improvement among Never Voted... that is just 138 people, essentially nothing.

Code:

NJ --- All Voters  ---           55.9% D --- 36.7% R --- 7.5% UNK ---                          D+6.1 Swing in Partisanship Margin from 2014
NJ --- Super Voter ---           55.9% D --- 41.5% R --- 2.6% UNK ---                          D+5.4 Swing in Partisanship Margin from 2014 ---              (R+0.3 Change in Swing Since Oct 31)
NJ --- Freq. Voter ---           55.4% D --- 36.4% R --- 8.2% UNK ---                          D+3.8 Swing in Partisanship Margin from 2014 ---              (R+1 Change in Swing Since Oct 31)
NJ --- InFreq. Vt. ---           56.5% D --- 30.1% R --- 13.4% UNK ---                          D+6.5 Swing in Partisanship Margin from 2014 ---              (R+2.1 Change in Swing Since Oct 31)
NJ --- Never Voted ---           56.6% D --- 27.6% R --- 15.9% UNK ---                          R+9.8 Swing in Partisanship Margin from 2014 ---              (D+2.8 Change in Swing Since Oct 31)

NJ doesn't have that much early voting, but what they do have looks very good for Dems. Bring on the Menendezslide! #TossupNJ

The only slightly worrying thing about NJ is R improvement among Never Voted. But still, Dems lead 56.6-27.6 in modeled partisanship among new voters, so...

I would not like to be a Suburban GOP incumbent in New Jersey right now.

Code:

NV --- All Voters  ---           48.0% D --- 44.9% R --- 7.1% UNK ---                          D+11 Swing in Partisanship Margin from 2014
NV --- Super Voter ---           44.1% D --- 51.4% R --- 4.5% UNK ---                          D+4.4 Swing in Partisanship Margin from 2014 ---              (R+4 Change in Swing Since Oct 31)
NV --- Freq. Voter ---           48.9% D --- 43.6% R --- 7.4% UNK ---                          D+11 Swing in Partisanship Margin from 2014 ---              (R+5.6 Change in Swing Since Oct 31)
NV --- InFreq. Vt. ---           53.3% D --- 37.6% R --- 9.1% UNK ---                          D+18.8 Swing in Partisanship Margin from 2014 ---              (R+4 Change in Swing Since Oct 31)
NV --- Never Voted ---           47.8% D --- 42.6% R --- 9.5% UNK ---                          D+8.6 Swing in Partisanship Margin from 2014 ---              (R+7.1 Change in Swing Since Oct 31)

Nevada looks great for Dems. Not only is there a large overall improvement for Dems from 2014 and a Dem edge overall in partisanship, but the Dem improvements and advantages are concentrated among unlikely voters (an important point that makes the #s look better for Dems than if you just look at party registration alone). Republicans are at more risk of cannibalizing their election day vote.

One interesting thing, that might give at least a bit of caution, is all the large R+ numbers in Swing Since Oct 31. But that just means that Dems improved over the last few days of 2018 less than they improved of 2014. I think that is more an indication of how horrendously bad the first week or so of early voting was in 2014 for Dems, as opposed to an indication of anything about 2018.

Unless we hear that nobody is voting on election day in Clark County, I would be very surprised if Rosen does not win.

Code:

OH --- All Voters  ---           39.6% D --- 48.6% R --- 11.8% UNK ---                          R+2.3 Swing in Partisanship Margin from 2014
OH --- Super Voter ---           40.7% D --- 50.7% R --- 8.6% UNK ---                          R+3.9 Swing in Partisanship Margin from 2014 ---              (R+3.1 Change in Swing Since Oct 31)
OH --- Freq. Voter ---           40.3% D --- 48.7% R --- 11.0% UNK ---                          R+3.1 Swing in Partisanship Margin from 2014 ---              (R+2.1 Change in Swing Since Oct 31)
OH --- InFreq. Vt. ---           36.2% D --- 46.2% R --- 17.6% UNK ---                          D+2.7 Swing in Partisanship Margin from 2014 ---              (D+0.2 Change in Swing Since Oct 31)
OH --- Never Voted ---           34.1% D --- 39.0% R --- 26.9% UNK ---                          D+9 Swing in Partisanship Margin from 2014 ---              (R+0.7 Change in Swing Since Oct 31)

Like Michigan, the early vote in Ohio looks surprisingly good for Republicans. This is one of the very few states with an R+ overall swing (R+2.3). On the bright side, Dems have at least improved with Infrequent and New voters a bit. But this does not make me particularly optimistic about Richard Cordray in the Gubernatorial race... Like GA and MI, this is one place where the early vote potentially tells a bit of a different story than the polls.

We don't know exactly how this translates into OH-01 and OH-12, but it also doesn't make me particularly optimistic about either of those (which I wasn't anyway). I am afraid Dems may have a rougher time in OH than in many other states.

Code:

OR --- All Voters  ---           54.7% D --- 38.9% R --- 6.4% UNK ---                          D+4.6 Swing in Partisanship Margin from 2014
OR --- Super Voter ---           55.6% D --- 40.9% R --- 3.5% UNK ---                          D+4.2 Swing in Partisanship Margin from 2014 ---              (R+2.5 Change in Swing Since Oct 31)
OR --- Freq. Voter ---           53.7% D --- 39.8% R --- 6.5% UNK ---                          D+6.6 Swing in Partisanship Margin from 2014 ---              (R+0.1 Change in Swing Since Oct 31)
OR --- InFreq. Vt. ---           54.9% D --- 35.0% R --- 10.1% UNK ---                          D+2.7 Swing in Partisanship Margin from 2014 ---              (D+1.9 Change in Swing Since Oct 31)
OR --- Never Voted ---           54.4% D --- 30.4% R --- 15.1% UNK ---                          R+5.4 Swing in Partisanship Margin from 2014 ---              (D+2.3 Change in Swing Since Oct 31)

OR looks solid for Dems, with a slightly above-average overall swing to Dems. I'd say this means RIP Knute Buehler.

Code:

TN --- All Voters  ---           29.0% D --- 63.0% R --- 8.0% UNK ---                          R+7.5 Swing in Partisanship Margin from 2014
TN --- Super Voter ---           27.3% D --- 68.2% R --- 4.5% UNK ---                          R+10.4 Swing in Partisanship Margin from 2014 ---              (R+1.4 Change in Swing Since Oct 31)
TN --- Freq. Voter ---           30.6% D --- 62.4% R --- 7.0% UNK ---                          R+8.9 Swing in Partisanship Margin from 2014 ---              (R+2.1 Change in Swing Since Oct 31)
TN --- InFreq. Vt. ---           30.0% D --- 56.7% R --- 13.3% UNK ---                          R+10.4 Swing in Partisanship Margin from 2014 ---              (R+2.4 Change in Swing Since Oct 31)
TN --- Never Voted ---           24.2% D --- 54.2% R --- 21.6% UNK ---                          R+12.7 Swing in Partisanship Margin from 2014 ---              (R+5 Change in Swing Since Oct 31)

TN looks bad for Dems. The only optimistic take for Bredesen is that hopefully a lot of the people who are modeled as Rs are voting for him. And indeed, for him to have any chance whatsoever, that would have to be the case anyway. Maybe the types of modeled Republicans who are voting are Bredesen Republicans. Otherwise, good for Blackburn.

Code:

TX --- All Voters  ---           38.8% D --- 52.5% R --- 8.7% UNK ---                          D+6.4 Swing in Partisanship Margin from 2014
TX --- Super Voter ---           35.0% D --- 61.5% R --- 3.5% UNK ---                          D+1.5 Swing in Partisanship Margin from 2014 ---              (R+2 Change in Swing Since Oct 31)
TX --- Freq. Voter ---           40.7% D --- 51.7% R --- 7.5% UNK ---                          D+3.8 Swing in Partisanship Margin from 2014 ---              (R+0.3 Change in Swing Since Oct 31)
TX --- InFreq. Vt. ---           39.9% D --- 44.3% R --- 15.8% UNK ---                          D+5.5 Swing in Partisanship Margin from 2014 ---              (R+1.3 Change in Swing Since Oct 31)
TX --- Never Voted ---           41.3% D --- 38.5% R --- 20.2% UNK ---                          D+9.3 Swing in Partisanship Margin from 2014 ---              (R+0.5 Change in Swing Since Oct 31)

Because of my special interest in TX I will say a bit more, and speculate about Beto's chances of winning...

First of all, TX is the main state where I wish we had 2016 numbers for comparison. They would be more informative than 2014. But alas... We have to do with what TargetSmart gives us here...

The early vote in TX DOES look very good by historical standards, in comparison to 2014. Youth turnout IS way up. Hispanic turnout IS is amazing compared to historical standards for a midterm, and compared to all reasonable expectations. Beto HAS run an amazing campaign, and HAS achieved something special, unique, and important for the future of the TX Democratic party and the National Democratic Party. Everything that people say to that effect IS true.

There IS a significantly larger than average swing (D+6.4) in the early vote in TX. And what is better, the gains are (somewhat) concentrated among Infrequent voters (D+5.5) and especially New voters (D+9.3).

Something special IS happening in Texas.

And nevertheless, after all this, there is an overall 52.5% - 38.8% Republican advantage (13.7% margin) in modeled partisanship among all early voters. As another way to look at that, even if you were to say that 100% of the people with Unclear partisanship were actually Dems, there would still be a 52.5% - 47.5% Republican advantage. The problem is that Democrats lost by 20 points in 2014. So not just a large improvement is needed, but a massive improvement is needed.

And here is the part that is worse news for Beto... Despite the large swings to Dems, concentrated among Infrequent Voters, there remains a Modeled R advantage even among Infrequent Voters (44.3%-39.9%). But to win, Beto doesn't need to just come close with infrequent/new voters, but to win them overwhelmingly, and turn them out in large numbers. It looks like there are too many modeled Rs among them for that to really be the case - or at least for it to be the case enough for him to actually win.

So despite how amazing the TX early vote is as compared to 2014, the data does NOT really suggest a Beto win. True, it suggests definite large over-performance as compared to 2014 (which I think will pay dividends in TX-07, TX-32, and hopefully a variety of other state legislative races and maybe even one or two additional congressional races if we are lucky), but Dems lost by 20+ statewide in 2014.

Somewhat similarly to Georgia and Tennessee, the only way the early vote data suggests the possibility of a Beto win is if there are really large numbers of people showing up in the data as Modeled Republicans, but who have abandoned the Republican party. To some extent that is probably true, but while polling indicates Beto is doing well with independents, it does not indicate that he is really doing particularly well with actual Republicans. So I am skeptical, and I would say to have any chance of winning, Beto needs a truly unprecedented and amazing election day. Beto voters and Democrats have to vote overwhelmingly on election day, and Republicans have to not keep voting like they have been voting in early voting so far.

While it is possible Beto could keep it pretty close, we should also not be too surprised if Cruz ends up winning by double digits (indeed, if Beto only loses by low double digits in a midterm, that would be a remarkable achievement and improvement by realistic standards for a TX Dem). I think that is a bit less likely than it was a week or so ago (the last week of early voting has been good), but it still remains much more possible than a lot of people seem to recognize.

The R+ swings since Oct 31 are reminiscent of Nevada, and I think potentially for the same basic reasons - they may indicate more about how horrendous 2014 was for TX Dems as opposed to how good or bad 2018 will be. But on the other hand, to have any chance of Dems winning, a much larger improvement is required in TX than NV.

Code:

UT --- All Voters  ---           27.9% D --- 54.8% R --- 17.3% UNK ---                          D+9.2 Swing in Partisanship Margin from 2014
UT --- Super Voter ---           19.4% D --- 65.2% R --- 15.3% UNK ---                          R+4.6 Swing in Partisanship Margin from 2014 ---              (R+3.3 Change in Swing Since Oct 31)
UT --- Freq. Voter ---           28.9% D --- 53.3% R --- 17.8% UNK ---                          D+8.2 Swing in Partisanship Margin from 2014 ---              (R+4.8 Change in Swing Since Oct 31)
UT --- InFreq. Vt. ---           38.6% D --- 42.1% R --- 19.3% UNK ---                          D+21.6 Swing in Partisanship Margin from 2014 ---              (R+5.5 Change in Swing Since Oct 31)
UT --- Never Voted ---           47.0% D --- 32.7% R --- 20.3% UNK ---                          D+39.8 Swing in Partisanship Margin from 2014 ---              (D+3.5 Change in Swing Since Oct 31)

Utah still looks amazingly fabulous for Dems (as compared to 2014), although it has gotten worse compared to 2014 over the past few days - I guess at some point gravity has to exert itself.

I am very optimistic about UT-04 given these numbers (although I wish we had them on the Congressional district level to tell that it looks like this there in particular).

Code:

VA --- All Voters  ---           49.3% D --- 38.9% R --- 11.8% UNK ---                          D+13.6 Swing in Partisanship Margin from 2014
VA --- Super Voter ---           48.7% D --- 43.0% R --- 8.3% UNK ---                          D+9.4 Swing in Partisanship Margin from 2014 ---              (R+0.8 Change in Swing Since Oct 31)
VA --- Freq. Voter ---           49.2% D --- 39.2% R --- 11.5% UNK ---                          D+14.9 Swing in Partisanship Margin from 2014 ---              (R+0.7 Change in Swing Since Oct 31)
VA --- InFreq. Vt. ---           51.8% D --- 31.6% R --- 16.6% UNK ---                          D+17.7 Swing in Partisanship Margin from 2014 ---              (D+2.5 Change in Swing Since Oct 31)
VA --- Never Voted ---           45.2% D --- 30.5% R --- 24.4% UNK ---                          D+9.6 Swing in Partisanship Margin from 2014 ---              (D+4.1 Change in Swing Since Oct 31)

VA early vote looks fabulous for Dems. We don't know exactly how it translates to the CD level, but it does not look like good gnus for Republicans in VA-02, VA-05, VA-07, and VA-10.

Code:

WA --- All Voters  ---           53.5% D --- 36.7% R --- 9.8% UNK ---                          D+6.2 Swing in Partisanship Margin from 2014
WA --- Super Voter ---           53.8% D --- 41.0% R --- 5.1% UNK ---                          D+2.6 Swing in Partisanship Margin from 2014 ---              (R+0 Change in Swing Since Oct 31)
WA --- Freq. Voter ---           53.6% D --- 36.3% R --- 10.1% UNK ---                          D+9 Swing in Partisanship Margin from 2014 ---              (R+0.2 Change in Swing Since Oct 31)
WA --- InFreq. Vt. ---           53.5% D --- 30.5% R --- 16.0% UNK ---                          D+5.4 Swing in Partisanship Margin from 2014 ---              (R+1 Change in Swing Since Oct 31)
WA --- Never Voted ---           48.7% D --- 28.1% R --- 23.2% UNK ---                          R+1.3 Swing in Partisanship Margin from 2014 ---              (R+4.5 Change in Swing Since Oct 31)

WA looks similar to OR for Dems - quite solid, with Rs potentially doing a tad better with late returned ballots than usual. The question is how that translates on the district level to WA-03, WA-05, and WA-08.

Code:

WI --- All Voters  ---           36.1% D --- 43.5% R --- 20.4% UNK ---                          D+2.1 Swing in Partisanship Margin from 2014
WI --- Super Voter ---           35.1% D --- 45.1% R --- 19.9% UNK ---                          R+2.5 Swing in Partisanship Margin from 2014 ---              (R+3.2 Change in Swing Since Oct 31)
WI --- Freq. Voter ---           32.9% D --- 46.2% R --- 20.9% UNK ---                          D+0.8 Swing in Partisanship Margin from 2014 ---              (R+2.7 Change in Swing Since Oct 31)
WI --- InFreq. Vt. ---           45.4% D --- 34.1% R --- 20.5% UNK ---                          D+17.2 Swing in Partisanship Margin from 2014 ---              (R+0.9 Change in Swing Since Oct 31)
WI --- Never Voted ---           46.7% D --- 31.8% R --- 21.4% UNK ---                          R+0.8 Swing in Partisanship Margin from 2014 ---              (D+3.3 Change in Swing Since Oct 31)

WI looks pretty good for Dems, despite a below average overall swing. The Dem gains are tremendously concentrated among Infrequent Voters (D+17.2), while Rs have improved among Super Voters (R+2.5), meaning R cannibalization risk. Although I might be a bit disappointed that the size of the swing is not larger, I would be very happy with the distribution of the swing and how if I were Evers, and how strongly it is concentrated among Infrequent voters.

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

Don't know why they bother trying to interpret EV in states where the rules are strict.

True it is more meaningful in some states than others, but you can still compare to the past in states where it is restricted. If there is a Dem improvement with early voting, that suggests a Dem improvement overall (to some degree, though of course election day can always be different).

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

TargetSmart EAV swing, day before 2014 to today:

Yup, OH and MI really stand out there.

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

Why are we still using TargetSmart's "modeled partisanship" junk? They thought Clinton would win Iowa and Ohio based on that, and we all know how that turned out.

Maybe someone's interpretation of the data thought that, but the data certainly didn't think or say that. If you have some sort of analysis that you think invalidates using their data, feel free to explain or link to it.

It is just voter file data. How you choose to interpret it is up to you.

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

Some news from my county!  Looks like early voting was insane this year.  It feels awesome to be among the 38,583!

https://www.nwitimes.com/news/local/govt-and-politics/elections/early-voters-shatter-the-old-mark-for-midterm-elections-across/article_8bad2801-1bcf-5446-a4b8-c80a35fec014.html

You must be logged in to read this quote.

Does anyone besides me think this bodes well for Joe Donnelly, assuming turnout in these areas is high tomorrow as well?

You must be logged in to read this quote.

Sounds promising, but do you have any idea how it compares to the rest of the state and to 2016? It is obviously more of a big deal if the increase is greater than elsewhere.

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

Don't forget Abbott has been averaging about 12 points above Cruz in polling. If you take all the TargetSmart data, assume that Unclear are 65-35 Dem, assume all Modeled R vote Abbott and then take 6% of the votes away and give them to Beto for Senate it results in <1% Beto win.

But Abbott is polling better than Beto primarily because he is doing better with Independents. Presuming that Unclear and Independents are at least roughly similarish sorts of groups of people, the way you are calculating it seems like it double counts the independents. Because you start by giving Dems most of the Unclear, and then also take into account Beto's better performance among independents again when converting from Abbott to Cruz. Whereas if you want to start from Abbott rather than the Senate race, you should start by giving Abbott most of the Unclears...

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

Cruz, you missed out one very important Race off the TargetSmart Early Vote Data and that's Indiana.

It's 56-35 GOP in the Early Vote. Republicans should feel very confident about that one. Actually HoweyPolitics which is well regarded in the Hoosier State where former Rep. Mark Souder is one of the Analysts predicting a narrow Victory for Braun AND if that happens it will effectivly end the Democrats Chances of retaking the Senate. Souder predicted that Donnelly will get Crossover & Indie Support but it won't be enough to topple Braun.

Oops, you are right, I should have had Indiana - I am not sure why I didn't earlier (probably either they didn't have many votes when I first set it up, or maybe they started EV later). It is about 500k votes, so that is actually a considerable amount, too.

Here are the #s for Indiana in the same format as the others I posted:

Code:

IN --- All Voters  ---           35.1% D --- 55.6% R --- 9.3% UNK ---                          D+7.6 Swing in Partisanship Margin from 2014
IN --- Super Voter ---           34.4% D --- 58.6% R --- 6.9% UNK ---                          D+5.7 Swing in Partisanship Margin from 2014
IN --- Freq. Voter ---           37.7% D --- 54.2% R --- 8.1% UNK ---                          D+11.1 Swing in Partisanship Margin from 2014
IN --- Infreq. Vt. ---           31.3% D --- 53.1% R --- 15.6% UNK ---                          D+3 Swing in Partisanship Margin from 2014
IN --- Never Voted ---           26.2% D --- 49.9% R --- 23.9% UNK ---                          R+6.2 Swing in Partisanship Margin from 2014

You have the overall partisanship margin there. That is good for Rs, but on the other hand it may be more or less in line with what one would expect for IN (and hence baked into the polling). It actually seems to look better for Rs when you break it down by vote history, as opposed to just looking at the top line.

The good thing for Dems is that there is a large swing (D+7.6) in modeled partisanship of early voters as compared to 2014, but it appears that the swing was concentrated among Frequent voters (and somewhat also Super voters, but in particular Frequent voters). There is a much weaker D+3 swing among Infrequent voters, and actually an R+6.2 swing against brand new voters (although that group is only 18,870 people, but in most states Dems are getting a swing in turnout of new voters).

So there should be some concern among Dems about cannibalization of the election day vote, and Dems in Indiana appear to be under-performing somewhat in turning out infrequent and new voters. But at least since the swing to Dems is more concentrated among Frequent than Super voters, the fear of cannibalization should not be quite as great as if it were more concentrated among Super voters in particular.

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

Cruz, you missed out one very important Race off the TargetSmart Early Vote Data and that's Indiana.

It's 56-35 GOP in the Early Vote. Republicans should feel very confident about that one. Actually HoweyPolitics which is well regarded in the Hoosier State where former Rep. Mark Souder is one of the Analysts predicting a narrow Victory for Braun AND if that happens it will effectivly end the Democrats Chances of retaking the Senate. Souder predicted that Donnelly will get Crossover & Indie Support but it won't be enough to topple Braun.

In a state like Indiana, Donnelly is going to get crossover support. So going by partisan EV is not most accurate. Also, there's only 9% indies in the EV? That seems low.

I think it goes by voting in past primaries if no party registration which is dubious.

I can't speak to the TargetSmart partisanship modeling in particular, but other partisanship modeling in Voter Files which is made in a similar way is usually based on a variety of factors (basically anything that can statistically predict partisanship), including primary voting, but also many other things such as demographics of age/gender/race/modeled race, but also data such as past election results in the precinct the voter lives in, census data about the census block the voter lives in, and also other things like consumer data. In general the modeling would be much less complete (with many more unknowns) if all that were used was primary voting history.

Indeed, the modeling is not going to be perfect, but it should be at least fairly close to the best that is achievable. You can criticize it by saying that it could be wrong (indeed, it could be and no doubt to some extent is), but the alternative is really just to have no data at all about the partisanship of the early voters. And I don't think that is a better alternative. It should generally be more accurate on a macro level over large numbers of individuals than for any one individual voter, at least if it is like other partisanship modeling.

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

I also think 2014 is a very bad year to compare to, given we didn't have any significantly contested races statewide.  All the congressional races were uncontested.

Similar problem (but in reverse) as to North Carolina then. Although it *usually* does make more sense to compare to the last midterm, It would be nice if they had 2016 data for comparison as well.