The Polls could be off and Democrats could win it all...from 538. It could go the the other way too

[Demsrule86]

of course but enough folks here are breathlessly posting bad news multiple times a day (why?) so I will leave a link for those who wish to explore that possibility or someone else can post because I won't. Safe to say this...no one has any idea what will happen on Tuesday In fact, all this premature screaming about election fraud makes me wonder if some on the Right know or suspect they will lose on Tuesday...here is hoping.






https://fivethirtyeight.com/features/2022-polling-error/

[onecaliberal]

1. 538 polls are not reliable. Joy Reid explained why on her show a few days ago.

[yardwork]

3. My understanding is that 538 isn't a poll. It's an aggregator of polls.

Garbage in, garbage out. 538's predictions are only as good as the accuracy of the polls they aggregate.

[onecaliberal]

4. Right wing skewed polls.

[BlueCheeseAgain]

33. They aggregate all polls...

... and try to account for pollster bias. Whether they do so successfully is hard to say. But I do think they're aware of the problem.

[Demsrule86]

9. That is why Nate is basically saying...hey these are the numbers for the Midterm...but if the polls

are off, it could be a very different outcome. We know that more than half the polls used for the aggregate are cheap shitty GOP polls. It is even possible the right knows they could lose and are going to scream 'stolen election based on the shitty polls.

[Amishman]

21. He's saying what I've said a lot - this is a particularly difficult election to model

and every pollster is dependent on their individual turnout model to align their actual sample to the expected electorate.

Models are tricky this time around because of the wild swings and surprises the past few cycles, as well as the impact of the pandemic on how and when people vote. It makes the most important data for modeling - recent results - less comparable to what is happening now.

[Demsrule86]

49. Honestly, I have no idea what will happen on Tuesday... I keep thinking we may win but who

knows? I find it interesting that the GOP is already crying stolen election BS.

[Deminpenn]

20. 538 does no polling, it's not really a poll aggregator either

What it does do is take polling data, which is massages to some degree, then runs 1000s of computer simulations, then adds up the outcomes and divides each by the number of simulations run. For ex, if it runs 100 simulations and those produce 51 outcomes where Dems hold a majority of senate seats. That translates into Dems having a 51% chance of winning/holding the senate.

538 is run by statisticians, not pollsters or political analysts.

[themaguffin]

32. The context is still correct, "538 polls" are not reliable, as in 538 includes shady R polls skewing

...the results.

[onecaliberal]

36. 👆🏻👆🏻👆🏻

It’s done purposely also. They know they’re doing it.

[Deminpenn]

39. Again, not defending 538, but they do make an attempt

to correct for partisan bias in the polls. I suppose it's an open question how well their corrections actually correct bias.

[onecaliberal]

45. If you can find the Joy Reid show from a few days ago, they had a guest on that explained

exactly why 538 should not be relied upon. It wasn't just them singled out either.

[Deminpenn]

38. 538 tries to correct for that bias, but still it has the GIGO

problem that all computer modeling has.

[yardwork]

54. Thanks for that explanation! That makes sense.

[Genki Hikari]

56. Um... Pollsters are statisticians. The real ones, no less.

Statistics is nothing more than the analysis of large-scale numerical data to infer proportions to a representative whole--which is exactly what pollsters do, day in, day out. That's why pollsters use the exact same statistical analysis formulae and mathematical processes as "other" statisticians.

Pollsters conform to the true definition of statisticians, because they are the ones who do the hard work of collecting and analysing data from representative samples, while 538 does...exactly NOT that.

In fact, 538 isn't doing statistics at all. They conduct no data collection of their own, and they test no representative samples of their own. So how can they do the crucial aspect of inferring proportions to a representative whole if they don't do the minimum of putting together a representative sample and collect data from that to analyse? They can't.

To put it bluntly, what they're actually doing is a meta-analysis of the statistical analyses of real statisticians, and, worse still, based on some questionable models.

[RussBLib]

2. ignore the polls and VOTE

like our lives depended on it.

Democracy sure does!

[Demsrule86]

10. Exactly...

Let's show and destroy the GOP. I think we can.

[ColinC]

5. It isn't guaranteed but I think a polling error favoring Dems is far

More likely this year than it has been for a very long time.

That being said knock on doors and vote

[Demsrule86]

12. With Roe and Medicare...I do too...time will tell!

[ColinC]

13. Early numbers, for the most part seem to be looking cgood

But we are going to need a really strong EDay GOTV push, i think. As well as the one happening now through Election Day.

[Scrivener7]

6. We will know when we know. I'm imagining a LOT of women

who will vote D but won't talk about it in front of their maga husbands. I really hope Kansas was a bellweather.

But we all process information differently. And we will know when we know.

[Demsrule86]

18. I think you are right...there will red areas that go for the Democrats...they have missed the Roe

polling completely...on election night the trend should become pretty clear after the polls close. It could go either way. But I am sick of the post that just insists we are doomed...maybe -maybe not but I still think we could win this.

[Scrivener7]

30. Some people like to believe their preferred outcome

is the likely one until and unless they are proven wrong.

Others like to prepare themselves for a bad possibility and are pleasantly surprised when things go well.

Two personality types. Neither right, neither wrong.

All Democrats.

[moonscape]

41. This!

[Demsrule86]

44. I have done GOTV forever and if people think we will lose that is when they don't turn up...

people read this site who may be lurkers or infrequent posters. Some folks post at rightie sites some of our posts. I don't doubt that people are worried. I am worried but I want to do everything I can to win.

[wryter2000]

26. Kansas was a real eye opener

No one could have predicted that. In CA, we have an amendment to add abortion rights to our constitution. That should bring out pro women's rights voters. There are some house seats up for question in the state. We have red areas.

[prodigitalson]

7. The conclusion of this analysis seems to be

'Who the fuck knows?'

We have lots of data that says you may as well flip a coin.

[Scrivener7]

31. Exactly.

[Polybius]

8. Well, there's only three possibilities

1) The polls are off in our favor, and Democrats will do better than expected
2) The polls are off in their favor, and Republicans will do better than expected
3) The polls are correct

You believe it's option 1, which would be awesome (and cause me to never doubt you again). I believe it's option 3. No wave, put Republicans take the House with 20 pickups and Senate with 1 seat. Let's all pray that it's not option 2.

[Demsrule86]

19. Ah but you missing an important fact...more than half of polls are cheap shitty right wing polls lik

Trafalgar, Insider Advantage...Wick. What Nate is saying is prepare yourself because we don't know what the fuck is going to happen.

[Polybius]

25. If they are proven wrong again, I hope 538 and RCP banns them

Rasmussen is another one, but that's weird. It seems they always over-favor Presidents. They consistently had Trump in the high 40's always. Remember when Biden was averaging high 30's this Summer? They had him at 45.

[Yavin4]

11. 50% chance of rain.

You can never be wrong.

[Demsrule86]

22. I have to wonder if there is a bit CYA going on...he has to know that some of the polls being used

are complete right-wing jokes.

[wryter2000]

27. You can be wrong in CA

We had one calendar year when it rained once where I live.

[Baggies]

14. I think bookmarking is the key.

That way after the election you can confront the poster and ask them why they were so insistent on being wrong.

[EarlG]

15. This is some world-class hedging

Pretty funny. No matter which way it goes they'll be able to say, "See, we told you so."

A good reminder that the only poll that counts is the one on election day.

[Demsrule86]

23. That is what I think...I am not giving up...going to get the vote out...I am walking later and the

rest of the week to GOTV...hubs can't walk but will call or text...hopefully we have an upset. I love math and polling and some of those polls are garbage...many partisan polls this year mostly right-leaning ones...so anything can happen. I am not saying we win. The president's party almost always loses the midterms. I am hoping for 'unprecedented' on Tuesday. The polls could also serve to bolster the claims that we stole the election should we win.

[Deminpenn]

40. For sure

nt

[Beautiful Disaster]

16. Polls in just about every high-profiled race are so close that any movement can change things.

The only reason Republicans are favored is because they lead in more polls but their margins are so thin in Georgia, Nevada and Wisconsin that just a slight shift toward the Democrats could win them all.

The problem the Democrats are facing right now in the senate is Nevada. That's a state that I think everyone took for granted for months while focusing on Georgia and Pennsylvania. If the Democrats lose both Nevada and Georgia, they'll have to flip PA + one other Republican state (OH, WI or NC) doable but unlikely.

If they only lose one, they can win it back with PA but their majority doesn't change.

There's also New Hampshire where polls have shown Hassan trailing, again, by a very narrow margin.

Really, it's a toss-up. It will come down to turnout. This isn't a wave election where it's in the bag for the party out of power. It's going to be a game of margins and turnout.

[Demsrule86]

24. And the Roe effect to tip the elections in our favor.

[bullimiami]

17. In summary. Polls are useless now. So why is there so much time wasted on them?

[Elessar Zappa]

28. I just take all polls with a grain of salt, good or bad.

We just need to focus on turning out the vote at this point.

[Sympthsical]

29. It's close and could go either way

Is there a hidden Dobbs undercurrent and over-correct for their under-sampling of Republicans last time around?

Or are polls including Dobbs and they've fine-tuned to accurately reflect Republicans?

The answer isn't knowable until Tuesday (or a few days after).

I think polls are basically in the right ballpark. However, I'm not quite as down on the Senate as some places are. I think Fetterman lands at a comfortable 2-3%% and Georgia goes to run off, where everyone will fight tooth and nail over it. Nevada is my true, "I have no idea, but I don't feel great about it." It's too early to know about AZ with Victor dropping out. I'm still generally comfortable with Kelly.

I still say Republican House at 218-225 area. Too many competitive districts favor them. A national number for Congress is useless, since each race is so localized. With blue voters condensed and red voters spread out, even when we're a little ahead in the national sentiment we're behind in results. A lot of polls show them ahead nationally, so I think the House is probably the most certain thing in any of this.

But hey, maybe that's wrong too.

We'll just have to see what we see.

[Demsrule86]

43. I say they lose the house because of Roe...women know you need both houses to codify Roe.

That is my hope anyway.

[PortTack]

34. Not sure why they have Warnock with less of a chance of winning. All the polls that would be

Part of their aggregation have Warnock either ahead, slightly ahead or toss up.

[Demsrule86]

35. Non-partisan polls show Warnock winning but there many partisan shit polls

There may be a runoff or there may be a hidden Roe vote.

[budkin]

37. They are trying to cover their asses

Because they know they got taken by all the garbage polls Simon Rosenberg talked about. I haven't looked at their site once since I heard that.

[Demsrule86]

42. Is that it...biggest CYA ever...hmmm

[iemanja]

46. I'm tired of all the complaining about the polls

(I know the OP isn't doing this. I'm responding to some of the responses in this thread and many other posts throughout DU).

Either pay attention to them or not, but this claim that 538 is a right-wing conspiracy is absurd. 538 aggregates polls, and it has a statistical model for weighting them based on reliability. That is ALL it does. For ANY poll to have much significance, the results need to be twice the margin of error. Very few polls show those kinds of results since the margins of votes in this country is so narrow. That speaks to the limitations of polls. One can certainly choose to dismiss polls for that reason or because we've seen so much inaccuracy in recent elections. But to claim a polling aggregator is a right-wing plot is ridiculous. People didn't have problems with 538 in the past, but now we are supposed to believe the site is out to get Democrats. It's a transparent excuse.

The election results will be what they are. We will know on Tuesday or Wed. You'll find out then if the polls were right or wrong, but regardless of the outcomes, there is no basis to claim the entire polling industry--and aggregators like 538-- are out to get you.

What there is, legitimately, is a problem in the media about how they report polls. They underplay the importance of margins of error and pretend that the polls are absolutely predictive. The media, and with it the public, have little understanding of statistics (count me among them), so they greatly overestimate what small differences in polls mean. They do the same thing every election cycle, even when they mistakenly overstated Democrats election changes. They may be doing to the same with Republicans this year--or not--but that's because of how the media handles polls, not because they (including, supposedly, MSNBC) have suddenly become MAGA. It's because they always do the same thing with elections and NEVER reconsider their approach when proven wrong.

[Laura PourMeADrink]

48. +1. May be crystal balls but only crystal balls we've got :)

[Demsrule86]

50. I don't disagree with you. Nate doesn't run the polls but he is warning that despite best efforts

on his part and other sites probably if the polls are even a little bit off...it could change the entire election outcome...so three things can happen...I can't help but think just like New York 19 and Kansas, these pollsters are missing the Roe vote and that their assumption which they make year in and year out may be wrong this time. Fingers crossed. Nate says to be prepared for any of the scenarios I posted below.

We win the house and the senate (my favorite outcome)
We win either the house or the Senate
We lose both the house and the Senate.

[Flash953]

47. Virtual phone bank this sunday and monday.

https://events.democrats.org/event/447536/?followup_modal_context=organization_newsletter_custom_recommendations

[lees1975]

51. Since all polls are considered and factored, and the GOP has flooded the market with polls to skew

the averages, my guess is that the Democratic advantage may be as high as 5% above where the margins of error come in.

And they may have been accurate in 2018, though not nearly as they were in 2008, 2010, 2012 and 2014, they missed big in 2020. And they've missed well outside the margin of error in bellwether races over the past year.

[yankee87]

53. So true

The polls over represent MAGAts because they are the only ones who answer phones. No one under the age of 70 answers the phone if they don't know the number

[Genki Hikari]

57. Stop.

Please stop.

You clearly don't know how pollsters operate in today's connectivity environment.

1) Pollsters don't rely only on landlines. They haven't for nearly 2 decades now.

2) Pollsters don't rely on cold calls to aggregate their representative samples.

Really.

3) Pollsters know how to use these things called voice mail and text messages to get in touch with people reluctant to answer calls from unknown numbers. Because those methods get through to even the most assertive phone screener.

4) I know this will shock you, but pollsters are willing to wait for the people they contact to return their calls. And you know what? Lots of people do return those calls. In fact, a very high number return them, compared to the old cold-call days.

So it's both a lie and just plain dumb to think that pollsters aren't able to reach all kinds of people across the political spectrum.

Because they are.

REALLY.

[yankee87]

58. Bad Polls

Why do you try to justify the fact the Reich wing is flooding the data with garbage polls? It's been shown to be true. They are doing this to try to depress the vote.

[renate]

52. I just don't get how the GOP can run Herschel Walker without shame and maybe get away with it

Especially against a man with such grace, kindness, and intelligence as Raphael Warnock. It shouldn't even be a contest. Even if he weren't literally pathologically stupid, he's just completely unqualified. It's mindboggling.

I know we were never the greatest country on earth, but good lord. How far we have fallen.

[Demsrule86]

55. Yup...I know. Hershel Walker is mentally ill and not smart IMHO...and unqualified for the Senate.