Final 538 House and Senate predictions

[Polybius]

59% chance Republicans take the Senate (most since June 14th). 84% chance Republicans take the House. Egads.

https://projects.fivethirtyeight.com/2022-election-forecast/senate/

[Dave says]

1. 'Muerica is so f'd up!

[Botany]

2. FYI

https://www.democraticunderground.com/100217332695

[secondwind]

3. THANK YOU!!!!!!!!!!!!!!!!

[Polybius]

6. Nate Silver knows about their fake polls

He takes this into account when predicting.

[SheltieLover]

4. Not buying it.

[Demsrule86]

5. Not happening...538 knows they have shit righty polls in their agregates...

[Polybius]

7. See post #6

Nate knows this and takes it into account.

[Fiendish Thingy]

22. But not really- did he adjust the 538 model?

My understanding is 538 just tweaked some of the weightings of these polls, which seems to me to be a significant undersized response.

While it may have been too late and too costly to reconstruct the entire 538 model to account for the deliberate manipulation of the polling averages, that’s what is required under the circumstances.

Once it became clear the polling averages were being manipulated, Silver should have either a)excluded all the R sponsored polls from the averages or b) issued a statement declaring that the GOP was flooding the system with shitty polls, and that no valid projections could be made for this year’s elections. (I know ABC wouldn’t let him do that, as it would cost them money, but that would have been the honest thing to do).

Instead, he kept reporting manipulated averages based on shitty polls, and then wrote an article throwing up his hands and saying “who knows?!?! It could go either way!”

[emmaverybo]

36. I won't forget when Silver, on the basis of two small states' vote, predicted a Bernie sweep of

states for 2020 primary. He extrapolated and was just flinging those numbers around. He and other credible pollsters totally wrote off Biden before South Carolina. Accordingly, the press went all in for Bernie then Buttigieg. I lost faith in Silver then. He did not account for his predictions with qualifiers. Once he gets dug in, that's it.
Unfortunately, polling has an impact on behavior. I don't think so this time. Too much at stake for us.

[ItsjustMe]

8. The trolls are very busy today

[Polybius]

9. Trolls have low post counts

I've seen one today.

[Kingofalldems]

18. Some have thousands of posts.

[Polybius]

20. Trolls don't survive that long

I think that you may be confusing those who tend to be negative/realist with trolls. There's a big difference.

[W_HAMILTON]

25. Indeed.

Also, beware, since they are quick to report actual LOYAL Democrats.

[iemanja]

31. Is Earl G a troll to you too?

Because he affirmed anticipated Republica wins as part of historical trends.

[Fiendish Thingy]

10. 538's numbers are based on the flood of R polls in the past few weeks

Although Nate Silver has hedged and wrote an article saying anything could happen, he apparently couldn’t afford to adjust his model mid-cycle, and isn’t willing to admit or even acknowledge that he, and most other media outlets, have been played by the GOP’s manipulations.

Silver’s response was to shrug and say “the Dems/MSM could’ve paid for more polls too”, without acknowledging that the response rate has dropped to less than 1% (it was approx. 30% about 20 years ago).

[Polybius]

12. See post #6

Nate knows this and factors it in his predictions.

[Fiendish Thingy]

26. See post #22

Silver’s response was inadequate, and allowed the GOP’s manipulations to continue unimpeded.

Silver’s manipulated predictions will be read far and wide, but his apologetic article saying “it’s a crapshoot, nobody really knows who will win” will be read and reported by very few.

At least MSNBC had the integrity last night to acknowledge this manipulation and the fact that any predictions being made are not supported by solid polling evidence.

[iemanja]

32. He weighs polls according to reliability

The GOP polls are weighted less strongly, unless they have a history of accuracy.
We'll find out soon enough how accurate the polls are.

[Fiendish Thingy]

35. But that adjusted weighting was manipulated too

Even if the shitty R polls are given less weight, that fact that they outnumbered more reliable nonpartisan polls by 10 to 1 (or more) in the last few weeks means those polls still had an outsized impact on the averages.

Nate got played.

[iemanja]

37. Because you saw a piece on cable TV telling you so?

Or are you a qualified statistician?

[Fiendish Thingy]

39. I do have graduate level training in research analysis

But I don’t hold myself out as an expert on 538’s model.

Model or no model, the evidence is clear that the GOP flooded the system with shitty polls (less transparency for sure in most of these polls) in a deliberate effort to manipulate the polling averages and drive the narrative in their favor.

[iemanja]

41. Can you explain the evidence to me?

Which polls in particular, and how are they weighted incorrectly?

[Fiendish Thingy]

44. Go to 538, pick a race, and look at the dates and sources of the data

I’d recommend picking a high profile senate race or generic ballot polls to see the evidence most clearly.

Look at the polls released in the last two weeks of October: a flood (10-20 depending on the race)of GOP sponsored polls from Trafalgar, Rasmussen, Insider Advantage and other shadier outlets with less history. Compare their results with the results from the very few polls from the reputable, nonpartisan polling outlets. You won’t even need a calculator to see how the R polls skew the averages (going from memory, R polls showing Walker +5-6, when the non partisan polls show a range from Warnock +5 to Walker +1)

As I said, Silver adjusts the weighting of the R polls, but when they outnumber nonpartisan polls 10-1 or more, the only adjustment in weighting that would counter the manipulation of the average would be to exclude the R polls altogether.

[iemanja]

46. I went to the Walker / Warnock race

https://projects.fivethirtyeight.com/2022-election-forecast/senate/georgia/

The only poll of those you listed that is used in the average is Trafalger. However, the average is slight. It looks like none of them, or very few, are outside the margin of error. Silver, based on that poll average, predicts a 63% chance of Walker winning. To me, that doesn't mean Walker is going to win. I'd say it's pretty close to a toss-up because of how close the polls are, but then I'm also incredulous as to how people could want Walker, so that colors my perspective.

Also, I'm watching poker now, the World Series of Poker Europe. It's amazing how many times the statistics don't hold up.

Whether Silver's prediction turns out to be accurate is a somewhat different question from whether the average is overwhelmed by GOP partisan polls. Now, I don't know all the origins of the polls, but they don't include most of those you indicated. It's also entirely possible the polls are wrong. I still think 538 does their best to balance polls according to statistical methods.

[Fiendish Thingy]

47. To get the full view of the flood of polls, you have to drill down further than the first page

The first page of polling results only gives the first week of November, when more nonpartisan polls began to be released.

If you click/tap on “show more polls” a couple of times, you’ll see the late October polls, and can easily see the flood of R polls released in that period that impacted the averages and drove the narrative. It’s even worse in some other races.

There were multiple polls of just 550 LV’s (who knows how that was screened) that must have had MOE’s of +/-8% or more. Reputable nonpartisan polls tend to have sample sizes of 1500-2000 respondents to get a MOE of around +/-3%, but that takes a lot more time than 500 responses, hence fewer polls released when the shady R pollsters are cranking multiple polls. Utterly meaningless and useless, unless you know that MSM will just report and highlight the spread and who’s leading in the horce race, giving the illusion of momentum where there is none.

[iemanja]

48. There is one other of the polls you mentioned among the full array of polls

plus a number of B/C polls. 538 says they weigh the newest polls are greatly. This is not to say their predictions will hold, but it's looking like they weight the polls in ways that speak to your concerns.

If Daniel Negranu can lose with pocket 5s to Seng's pocket 4s (a hand that wins 4 our of 5 times) Democrats can prevail in tighter margins. Of course all in poker is pure odds and there are no human factors at that point.

[emmaverybo]

38. Thanks for your clarifying and thoughtful posts on this. nt.

[gab13by13]

11. I love it,

Fuck 538

POlybius, want to make a bet, for charity? 100 bucks we keep the Senate, 100 bucks we keep the House.

PM me if you are interested.

[Polybius]

14. I think there's an excellent chance we'd both be out $100

The Senate will be damn close, and it really could go either way.

[Ocelot II]

13. 538 gave Clinton a 71% chance of winning in 2016, so there's that.

[Polybius]

16. Super true

I think the Senate is a total toss-up, but not the House.

[RobinA]

23. Popular

or electoral? She did, in fact, win the popular. I'm with 528 on this. Some people are in for a biiiiig disappointment come the end of this week. Unfortunately.

[Peacetrain]

15. Silvers predicted Hillary would win..

just saying

[Polybius]

17. Yes he did, with a 71% chance

I think the Senate is a coinflip.

[iemanja]

33. And she had a majority of the votes

as most polls predicted.

[The_Casual_Observer]

19. This is a pretty half assed nothing prediction from those guys. Might as not even wasted their time.

All their never ending opinion bullshit about the historical red wave and they arrive at 50/50.

[RobinA]

24. I Think They're

hedging a bit with 50-50. Just in case. Red wave is more in keeping with the numbers no one wants to believe. We will see.

[underpants]

21. Hoping but not expecting great results.

Gas prices killed us.

[emmaverybo]

43. MSM did not bother to report widely enough that gas prices went down, down, down. Pre-inflation,

economy headlines read like this: "Economy booming under Biden, but..." and then ominous prediction about inflation or the pre-"but" clause was canceled out by people's concern over inflation signs. "...booming BUT folks scared to death that worst inflation in all of U.S. history..."
And so MSM take on the economy has gone with an emphasis on inflation, possible recession around the corner, and de-emphasizing of a) all the other signs--jobs, low UE, deficit lowering, GDP that may bode more permanent and telling signs of future economic growth.
Never a headline like "Although GOP hair afire 24/7 over inflation, they refuse to pass anti-gas price gouging bill..." " Under Biden, inflation is less high as compared to other countries in the grip of global ..."
MSM as usual is guiding political thinking and behavior. And as usual, news moved off of Roe vs. Wade overturn shocker and its fall out onto more GOP-friendly topics, inflation and crime.
I canceled my WaPo subscription. So-called liberal, they are one if the worst offenders.
If MAGAQ-GOP wins midterms in the face of their radical agenda, insane subjugatiion to Trump, and pro-Putin, Neo-Nazi leanings, I will blame WAPO and similarly "liberal" corporate-controlled news outlets almost as much as Fox.

[aeromanKC]

27. McMullin winning in Utah will screw Nate's numbers up.

[misanthrope]

28. Then I need to make out a last will and testament

Once they eradicate Social Security Disability and Medicare, I won't make it a full calendar year. I can't afford medical treatment or pharmaceuticals out of pocket.

[Efilroft Sul]

29. For what it's worth, let's factor in the weather.

Weather across the contiguous states has been generally good, which always helps with Democratic turnout. My only areas of concern where Democrats are viable are California and western Nevada receiving rain and snow, plus Minnesota and Wisconsin experiencing rain now and throughout the rest of the day.

Back to Katie in the studio.

[littlewolf]

30. I expect to lose the house.. that is fairly normal in a mid term

the Senate .. not sure ... prob not ...

[iemanja]

34. I've put the alcohol and mixer in the fridge

In anticipation for watching the results tonight.

[Polybius]

40. I'm getting drunk either way

Either to celebrate or sulk.

[Elessar Zappa]

45. Same.

I have a six pack of Budlight, which should do the job quite nicely.

[beaglelover]

42. 538: Hillary in a landslide! Nuff said about the cred of 538.

[Kingofalldems]

49. Oops---looks like a big GOP fail.