A Question For People Who Still Believe In Polls

[gab13by13]

Why is it that when the polls are wrong, they are always wrong in favoring the Magat candidate?

Do not believe in polls unless they are exit polls.

My poll, I predicted Suozzi by 5 will be closer than the supposed reliable polls.

[SoFlaBro]

1. What did the fucking polls say about the race?

[gab13by13]

3. The one I saw here, Suozzi +1

[SoFlaBro]

10. So a win well within the margin of fucking error.

[Mountainguy]

4. Most were Suozzi +4

[SoFlaBro]

11. It could jolly goddamn well end up being a double digit win.

[Mountainguy]

13. I don't think so

The early vote all has been counted, and election day votes are swinging pretty big +R. No shock, but it's already down to around an 11 point lead and that'll tighten as the ED votes tally.

There just aren't enough out there to overcome the lead Dems built up.

[Mountainguy]

2. Almost every poll I saw

Was Suozzi +4, so I'm not sure saying that +5 will be closer is a big win, even if it's true.

[gab13by13]

5. Did not see a +4 here at DU

and I don't click on polling sites.

[Mountainguy]

6. Here they are

https://projects.fivethirtyeight.com/polls/house/

[Melon]

7. I don't have a home phone, and don't

answer unknown calls on my cell. Who do they reach?

[Johnny2X2X]

8. It's over 17 points right now

So the polls were probably way off in favor of the Republican.

[Mountainguy]

14. I bet those +4 polls

end up pretty close. Even if they under estimate D support, some of that can probably be contributed to the storm depressing ED vote, which was heavily R leaning.

[Silent3]

9. Polls this year have missed a lot of Dem support. (Polls aren't as historically bad as a lot of people seem to think.)

Current polling models are always based on previous elections, and, for one thing, I don't think the way Dobbs has motivated Democratic voters has worked its way into how the models are weighted.

For another, perhaps despite all of their noise, Republicans just aren't as motivated as they have been in the past.

Response rate on polls has also gotten really, really low. Worse than I'd realized. Low response rate can be modeled and accounted for, but there are limits. I think we're getting into a lot of statistical noise now with so few people willing to respond to pollsters.

When it comes to aggregate polls, there are a lot of slanted Republican junk polls being thrown out there lately, which should be ignored, but are nevertheless getting into the mix and throwing off aggregate results.

[tableturner]

12. Dirty secret: polls are based on pollsters' assumptions about turnout. A wrong assumption gives a wrong conclusion!

Edited because I put an equals sign in the title, had to change it to "gives".

[Yavin4]

15. All political polls measure what people tell pollsters. Not who actually votes on election day.

All self-reporting polls should be taken with a high degree of skepticism. What wins elections is who turns out to vote. Trump's supporters are extremely enthusiastic about Trump but no one else.

Polls are not good at predicting results unless incumbency is involved or high name recognition.