Is the polling industry wrong, or is something else wrong?

[The Economy is Getting Worse]

All the pollings sugggest Trump is ahead right now, nationally and in the states.

But election results from last week suggests that Democrats are not dead.

What's going ON?

1.Democrats benefit from very low turnout ever since Dobbs. Special elections between the same 2 candidates tend to be 10 points bluer than the midterm result (see Nebraska-1).

2.Democrats are way outperforming at the Senate level, like in 2022 (the House level is more mixed, where Rs won the popular vote by 2.8 in 2022). Sherrod Brown is still up.

3.The results in Virgnia were not great for Democrats. They still did 8-9 points worse than Biden, which would give Trump a 4 point popular vote win if applied nationwide.

4.Some voters probably blame Biden for Dobbs. Biden also gets blamed for the bad economy more than downballot Democrats do.

[The Economy is Getting Worse]

Nonresponse bias was a major issue in 2020. It's not hard to believe it's an issue this year, just for the opposite side. One side has an active primary, the other doesn't. That can lead to whacky results, since one side is much more energized and paying attention to the race while another side isn't.

It's the same reason why some days we have Trump +1 with young voters and other days we have Biden +25. A lot of pollsters don't seem to be very good at working through this.

Another thing to note is that the generic ballot should be showing some strength for Democrats right now, given current election results. But it doesn't, it's basically 1 or 2% within the Biden/Trump margin.

We had an actual midterm in 2022, where the generic ballot was R +2.8. The generic ballot results make sense.

[The Economy is Getting Worse]

538 released an in-depth article explaining that the 2023 elections don't mean anything in reference to 2024: https://abcnews.go.com/538/2023-elections-us-2024/story?id=104916558

LMAO - that is not all what it says. It pretty much boils down to: who knows.

So after all that analysis, I'm afraid that we're not much closer to answering the question of what last week's elections mean for future ones. Some of 2023's election results clearly tell us nothing about 2024; others may tell us something, but we can't really be sure; still others have historically been quite useful but are pointing in confusing directions this year. In other words, I don't recommend changing your 2024 outlooks based on what happened last week.

Also, the fact that he started the article with this is just ... so wrong on so many levels:

But after a Republican sweep in the 2021 elections failed to predict a mixed-but-generally-good-for-Democrats 2022 midterm

As if there wasn't some major news event in the middle of those two things.

You seem to be a very black or white type of person. No one is saying Ohio is going to go blue or Biden +12 after last weeks election results. But there are clearly *trends* that are happening that cannot be ignored.

Democrats still lost the Generic Congressional Ballot by 2.8 in the 2022 midterms, which isn't far off where it is now. The Democratic Incumbents did outperform in the Senate in 2022, but polling also suggests that Sherrod Brown is favored right now, for example.

[The Economy is Getting Worse]

Nonresponse bias was a major issue in 2020. It's not hard to believe it's an issue this year, just for the opposite side. One side has an active primary, the other doesn't. That can lead to whacky results, since one side is much more energized and paying attention to the race while another side isn't.

It's the same reason why some days we have Trump +1 with young voters and other days we have Biden +25. A lot of pollsters don't seem to be very good at working through this.

Another thing to note is that the generic ballot should be showing some strength for Democrats right now, given current election results. But it doesn't, it's basically 1 or 2% within the Biden/Trump margin.

We had an actual midterm in 2022, where the generic ballot was R +2.8. The generic ballot results make sense.

2022 was not a national campaign, not sure why you're comparing the two. If you wanted to compare anything from 2022, you'd compare actual swing states/competitive races, where Democrats largely did well.

Democrats still did worse in the swing states in 2022 in the Generic Congressional Ballot, losing everything except for Michigan, where abortion was on the ballot. In 2024, higher turnout in the whiter battlegrounds (ie Wisconsin, Michigan, Pennsylvania) probably hurts Democrats, because the drop-off is mostly non-college white.

[The Economy is Getting Worse]

All the pollings sugggest Trump is ahead right now, nationally and in the states.

But election results from last week suggests that Democrats are not dead.

What's going ON?

Biden is not all Democrats.

Do you think every one of his polling ills can be traced back to his age? You think Americans give Democrats the thumbs up, but not Biden because he's too old for them?

Yes, it’s mostly his age.

But more generally, the Democratic candidate for Governor of Kentucky won by 5 points.  The Democratic candidate for Secretary of State of Kentucky lost by 21 points on the same day.  What’s going ON?

We are returning to the political norm in the Biden political era. The Trump era was an aberration with uniquely low split ticketing, no incumbency advantage, and uniform swings. We are now seeing more ancestral coalitions, higher split ticketing, and higher incumbency advantages.

[The Economy is Getting Worse]

538 released an in-depth article explaining that the 2023 elections don't mean anything in reference to 2024: https://abcnews.go.com/538/2023-elections-us-2024/story?id=104916558

LMAO - that is not all what it says. It pretty much boils down to: who knows.

And that means the article's point is that the  2023 elections can't be used as some sort of foreshadowing of what 2024 will be. Your take is, "He didn't say yes or no. So let me go ahead and lean towards yes"

Well, yes, they're still major elections, so it would be ridiculous to just ignore them as if they mean nothing. You can always gleam something out of them. To ignore an election that has millions of voters in favor of n=500 polls is asinine imo, but you do you.

Yes, that's why Romney won easily after Republicans did so well in 2010. I wonder why Nate Silver used n = 500 polls suggesting a small Obama advantage in the battleground states to forecast 2012 instead of actual election results from 2010.

[The Economy is Getting Worse]

The national polls are currently not reliable. If you look at each state polling, from Atlas polling aggregate, and distribute the weight of each state in the popular vote, Biden and Trump are basically tied at 41%, with Biden holding a very tiny lead if you add decimals. Of course, it's still early and a lot of states don't have strong trends to show, but Trump +10 or Biden +10 polls nationwide, are junk and not remotely close to the truth.

A lot of the state polling is old. The newer/higher quality state polling (ie NYT) shows Trump up by 5 in the swing states. Also, California/New York seem to be shifting hard right in state polling, as well as turnout differentials between red/blue states (ie blue states probably turned out more in 2020 due to mail-ins), could give Trump the PV.

[The Economy is Getting Worse]

Nonresponse bias was a major issue in 2020. It's not hard to believe it's an issue this year, just for the opposite side. One side has an active primary, the other doesn't. That can lead to whacky results, since one side is much more energized and paying attention to the race while another side isn't.

It's the same reason why some days we have Trump +1 with young voters and other days we have Biden +25. A lot of pollsters don't seem to be very good at working through this.

Another thing to note is that the generic ballot should be showing some strength for Democrats right now, given current election results. But it doesn't, it's basically 1 or 2% within the Biden/Trump margin.

We had an actual midterm in 2022, where the generic ballot was R +2.8. The generic ballot results make sense.

And that's why the 2010 R+6.8 led to President Romney in 2012 and 2018 D+8.6 led to democratic supermajorities in 2020.

Yeah I don’t think 2022 is reliable. I think Florida will trend left and Pennsylvania will trend right in 2024.