Election models megathread

[ProgressiveModerate]

My guess is the model has the GOP doing significantly better in the Senate than one expects, ah least relative to the House forecast.

Statistically, the senate and house predictions by pundits don’t rlly align because many pundits are treating races in Biden + 0/1 states as true tossups while Biden + 1 House seat is immediately written off as an R win. I suspect the model for instance will have a similar rating for Senate seats like AZ/GA/PA with House seats like IA-03/TX-15/VA-02

[ProgressiveModerate]

Honestly the Senate forecast doesn't seem too unreasonable and the House topline makes sense though there are def a few races that individually don't make much sense. They just don't align with eachother, though tbf overall they pretty much align with Atlas perhaps witha  slight boost to Dems.

The Governors forecast seems quite favorable to Dems in some places (i.e 91% Whitmer) but considering the fundamentals of the GOP candidates literally getting arrested, Whitmer's strong 2018 performance, and relatively good approval and poll numbers, I could see why a model would output that.

WI and KS are also a bit suprising; perhaps incumbency advantages and approvals skew things? After that, it seems relatively reasonable though def favourable to Dems

Making a statistical model is hard, trust me I've done it myself. The reason it trickly is because you know the polls could be slightly skewed (i.e the Warnock + 10 poll) but how you get a model to objectively be able to tell that poll is BS just becuase YOU think it is.

Their 2018 Senate and House forecasts were relatively decent all things considered and their 2020 model held up better than a lot of pundits though still slightly overestimated Ds. Even in 2020 House they still rated races like CA-25 and CA-21 as tossups even as many had written those off and generally didn't buy into the MASSIVE D House wave narative.

With the resources out there, the 538 model is as good as it gets, especially when a lot of the resources a model is reliant upon are flawed.

[ProgressiveModerate]

The senate model is clearly underestimating Republicans due to reliance on polls, the GOP has outperformed what the polls said in the Senate 4 cycles in a row, 2014, 2016, 2018 and 2020, in all 4 cycles the GOP won more senate seats than most projected, no reason to believe the polls have corrected for the bias.

I know that’s true for 3 of the elections, but are you sure about 2018? I thought the results matched the polls in the Senate then - it was only Democrat hopium that was defeated.

The finial 2018 FL Senate rcp average was was Nelson+2.4
The final rcp average for Indiana Senate 2018 was Donnelly +1.3, Braun won by 6

In 2018, polling on net was pretty accurate, but it underestimated partisanship which def lead to Dems being overestimated in a lot of key Senate races in deep red states. In most "normal" swing states polling was pretty good. (PA, WI, AZ, NV, TX).

We saw a simillar theme too in 2020 where a lot of polls showed very close Senate (and Pres) races in normally R states (SC, KS, AK, MT) that ended up being far off. However, a lot of the polling in deep blue states was pretty decent.

[ProgressiveModerate]

One thing this 538 Deluxe model seems to be good at dealing with is polls that just go very against fundamentals. For instance, in PA-Sen, the 4 most recent polls are Fetterman + 11, 14, 9, and 11. I don't think anyone in the election community expects him to win by that much, but these types of polls haven't pulled the model to declare the race like 95% Fetterman. We see a similar theme in OH where all these polls of Ryan leading haven't made OH-Sen race even a tossup and gives Vance a 3/4 shot.

One thing I don't understand is how the model has reached the conclusion that MI-Gov is Safe D despite basically no high-quality polls. That seems very overconfident, though I agree that fundementals prolly heavily favor Ds.

The other thing is that there seems to be a disconnect between the House forecast and the Senate/Governors forecast. The House Forecast expects an R + 4.5 margin nationally which isn't unreasonable, however, that doesn't align with Dems being favored in all these Senate and Governors races in narrow Biden states.

[ProgressiveModerate]

538 now gives Mark Kelly the same odds as Johnson and Budd at winning. (68-32)

As someone on the ground in AZ, does this seem about right to you or is it bullish on Kelly?

[ProgressiveModerate]

https://twitter.com/NateSilver538/status/1558095352485511170

I'm very confused about how the House model can think the GCB is slightly over 2020t R + 4 (about a 9 point swing from 2020) yet the Senate model can in good faith rate PA-Sen and AZ-Sen as Lean D. Sure, Oz and Masters have their issues but are they significant to matter that much?

And the Governors forecast almost seems to indicate a D wave, with MI-Gov being safe D, MN, WI, and PA being likely D, and KS being tossup.

Even just look at a state by state level in the House. AZ-Sen is lean D yet AZ-01 and AZ-06 are safe R, districts that Kelly will prolly have to win or at least come close in to win statewide.

I have a feeling that in reality their 3 models are not very well correlated at all.

[ProgressiveModerate]

https://twitter.com/NateSilver538/status/1558095352485511170

I'm very confused about how the House model can think the GCB is slightly over 2020t R + 4 (about a 9 point swing from 2020) yet the Senate model can in good faith rate PA-Sen and AZ-Sen as Lean D. Sure, Oz and Masters have their issues but are they significant to matter that much?

And the Governors forecast almost seems to indicate a D wave, with MI-Gov being safe D, MN, WI, and PA being likely D, and KS being tossup.

Even just look at a state by state level in the House. AZ-Sen is lean D yet AZ-01 and AZ-06 are safe R, districts that Kelly will prolly have to win or at least come close in to win statewide.

I have a feeling that in reality their 3 models are not very well correlated at all.

I do think the 538 model tends to overvalue incumbency across the board, which somewhat explains AZ.  Like I definitely don’t think Dems are more likely to win NC than WI. 

But candidate quality can easily create a >5 point swing in individual race; we see this in multiple races in every cycle.  Just in Pennsylvania alone, Bob Casey would have won in a R+4 environment in both 2012 and 2018 if you apply a uniform swing from the House PV.

To their credit, they had Maine Senate as a tossup basically through all of 2020 while many had written it off by ED.

[ProgressiveModerate]

Something really seem to shoot up Dems chances today at 538.

[ProgressiveModerate]

IDK, the DDHQ election model seems kind of sloppy. They seem to be missing quite a bit of polling in different races.

At least they fixed the MN-05 and FL-24 problems. Overall it doesn't seem too bad and seems like what you'd expect if 2022 ends up being a pretty solid R wave. My only gripe is that they weight incumbency a bit too mcuh in such a polarized era.

[ProgressiveModerate]

The Map isn't there for Democrats to retain the House.

I will be watching the Virginia 2nd Congressional Race on Election Night between Rep. Elaine Luria and her Republican Opponent Jen Kiggans. This District is a total Swing District similar to NY-19. If Kiggans wins Republicans Chances of retaking the House jumps to 95 %.

VA-02 was the median House seat on 2020 Pres numbers so yeah def a very important one for both sides (since it seems unlikely Kiggans or Luria would significantly defy traditionally partisanhip and run away with it). It seems like most paths of least resistance to a Dem majority will require them to win a few Trump districts such as IA-03, MI-07, PA-08, or MI-10 since Rs have seats like PA-01, CA-40, NE-02, and maybe NJ-07 pretty locked down (at least enough to be to the right of the tipping point)

For reference, he's a list of seats ranked by 2020 Pres results:

205 - WA08 - B + 6.67
206 - NV03 - B + 6.64
207 - NE02 - B + 6.32
208 - CA45 - B + 6.18
209 - NH01 - B + 5.93
210 - NM02 - B + 5.88
211 - PA17 - B + 5.83
212 - PA01 - B + 4.64
213 - NY19 - B + 4.62
214 - CO08 - B + 4.55
215 - KS03 - B + 4.45
216 - NJ07 - B + 3.66
217 - OH13 - B + 2.82
218 - VA02 - B + 2.05
219 - MI08 - B + 2.03
220 - CA40 - B + 1.87
221 - NC13 - B + 1.69
222 - AZ01 - B + 1.48
223 - MI07 - B + 0.95
224 - PA07 - B + 0.62
225 - NY01 - B + 0.21
226 - AZ06 - B + 0.07
227 - FL27 - T + 0.32
228 - IA03 - T + 0.34
229 - MI10 - T + 0.98
230 - CA41 - T + 1.09

One thing that's interesting is that seat 211 (PA-17) is nearly Biden + 6 which is pretty blue and to the left of the nation but after that the seats drop off for Dems very quickly and by seat 218 you're at Biden + 2. However, after that the list slows down again and you have quite a lot of seats that were extremely close in the 2020 pres race.

[ProgressiveModerate]

This is likely a rounding error but lol.

[ProgressiveModerate]

This is likely a rounding error but lol.

There’s a 1% chance that Wisconsin will just give a participation award to both of them.

Technically doesn't every race have like a 0.00001% of exactly tying. We've had some very close calls before

[ProgressiveModerate]

Will the 538 Deluxe model reach a point that D Senate chances are higher than R House chances?  (This is already true in the Classic model.)

The consistent trend towards Dems in the 538 model just comes from them holding their ballot lead as we get closer to the election, so it's quite possible.

538 model has been quite stubborn about moving the House model towards Ds much even as the Senate model zooms left. I think a lot of that has to be because Senate races get polling House races generally don't get. if Kelly really has a 76% chance of winning AZ, AZ-01 and AZ-06 are not likely almost solid R and idk why the model has such a hard time understanding that.

[ProgressiveModerate]

Will the 538 Deluxe model reach a point that D Senate chances are higher than R House chances?  (This is already true in the Classic model.)

The consistent trend towards Dems in the 538 model just comes from them holding their ballot lead as we get closer to the election, so it's quite possible.

538 model has been quite stubborn about moving the House model towards Ds much even as the Senate model zooms left. I think a lot of that has to be because Senate races get polling House races generally don't get. if Kelly really has a 76% chance of winning AZ, AZ-01 and AZ-06 are not likely almost solid R and idk why the model has such a hard time understanding that.

I mean they have Peltola at like 15.

I don't blame them there cause RCV is a lot harder to model.

[ProgressiveModerate]

Also, why the hell does the RCP model have the GOP winning Wisconsin and Pennsylvania when their polling averages have Fetterman and Barnes ~5 points up in each?

Because RCP is not a neutral site. It had an R-bias for quite some time, especially with the duration of including certain pollsters in polling averages.

In 2000, RCP had California as only Lean Gore before election day.

They had much bigger problems than that. They were projecting like a 400EV Bush win though tbf that was 20 years ago

Also in 2000 Cali was only like a D+10 state or so

[ProgressiveModerate]

Here are the current 538 polling averages in the competitive (including remotely competitive) states:

AZ D+7.4
CO D+9.1
FL R+3.8
GA D+2.1
NH D+7.5
NV D+0.7
NC R+0.3
OH D+0.6
PA D+6.5
UT R+7.0
WI R+1.0

I suspect that if you showed this to Democratic strategists back around March, they'd have been wildly ecstatic.

Kinda interesting how large the divergence is in the polling averages between Wisconsin and Pennsylvania.

In the end, I think it's pretty unlikely WI and PA Sen will vote that far apart from eachother, but polling in WI tends to be worse than PA to the point where polling underestimating Rs in WI feels like a rule.

It is worth noting that in 2018, both WI and PA polling were pretty spot on.

[ProgressiveModerate]

Here are the current 538 polling averages in the competitive (including remotely competitive) states:

AZ D+7.4
CO D+9.1
FL R+3.8
GA D+2.1
NH D+7.5
NV D+0.7
NC R+0.3
OH D+0.6
PA D+6.5
UT R+7.0
WI R+1.0

I suspect that if you showed this to Democratic strategists back around March, they'd have been wildly ecstatic.

Kinda interesting how large the divergence is in the polling averages between Wisconsin and Pennsylvania.

In the end, I think it's pretty unlikely WI and PA Sen will vote that far apart from eachother, but polling in WI tends to be worse than PA to the point where polling underestimating Rs in WI feels like a rule.

It is worth noting that in 2018, both WI and PA polling were pretty spot on.

And that's always a big asterisk with any polling analysis this year that might lend credence to the "Trump-on-the-ballot-effect" if it ends up being true.

If polls are relatively accurate similar to 2018, more or less, then Trump's presence really is a likely culprit of the misses in years' past. If they are as off as much as in 2020 then there really might be no hope for or reason to expect accuracy in American polling. This is kind of the fork in the road and the ultimate determining factor to me of what to expect from the polling industry from now on. And I don't how the inherent, near-universal disregard for most polls will play in future elections for both the campaigns and observers.

That said, there are still some states that were way off in 2018 and it should be expected for them to be that way this year and from here on out too. Namely Florida and Ohio. Maybe Nevada too in the other direction.

I don't think we've ever had a cycle where polling has been accurate EVERYWHERE. I think when we discuss polling error, we're asking nationally who do polls tend to underestimate and by how much. In 2020, it was def Rs even though in a few places like Colorado polls underestimate Biden slightly.

[ProgressiveModerate]

5 Reasons For Democrats To Still Be Concerned About The Midterms

https://fivethirtyeight.com/features/5-reasons-for-democrats-to-still-be-concerned-about-the-midterms/

The primary merit of this piece is that it is a revealing description of what is under the 538 hood, and what isn't.

Thanks Nate, but trust me, we're still fretting. No need to twist the knife more.

I've been in a nonstop state of anxiety about the country imploding for seven years now, and it's surprising how little the people who feel that way get acknowledged anywhere in the press.

I think oftentimes those who are genuinely very concerned rub off the wrong way in any sort of medias they do. Yes, we need to be aware of legitimate and serious threats to our democracy, but a lot of these people lack any optimism or hope as to how we can ensure those worst-case scenarios do not happen. Without that key piece, they're likely to be trampeled by the media and public alike cause people generally don't like full on dooming like that.

[ProgressiveModerate]

I'm a bit confused on the NC-SEN model in 538. We keep getting more and more tied polls of the race and yet the model has been drifting away from Beasley. If the race keeps being tied even until Election Day, why is it not moving - towards - Beasley a bit?

It seem NC is the most off from the polling average right now to the 538 forecast (tied in polls but Budd +3 final result)

Because for a bit Beasley lead for several polls in a row, plus as the election gets closer uncertainty tends to decrease.

[ProgressiveModerate]

Unfortunately Nate "Plastic" Silver has become a complete partisan hack.

My impression is that the Twitter consensus has come to believe Nate Silver is a Republican partisan hack due to some of his positions on covid.

Silver himself def seems to be somewhat of a moderate Democrat but also develops his own nuanced opinions on certain topics such as COVID. The people running any of these forecasting sites def have some degree of bias, the issue is how much they let it interfere. I tend to find Silver's models imperfect but pretty objective compared to RCP which basically sorts out the polls they dislike and also has way too many tossups so they can never be called out for wrong predictions.

[ProgressiveModerate]

https://twitter.com/NateSilver538/status/1584239922579324928

It feels weird that "Dems slightly overperform" the polls is totally off the table. Reason I say that is given the margins in the last couple of special elections. Yes, Dobbs isn't as close to the news as then...BUT the economy was objectively worse at that point compared to now, as well as a large part of the polling being wrong is that a segment who wasn't answering polls or otherwise politically active turned out. Think of it as a reverse Trump 2016.

Even relative to 2020, it seems like basically everyone is assuming Dems will underperform polling o some degree, even posters often considered “D hacks”