• Moderators, please send me a PM if you are unable to access mod permissions. Thanks, Habsy.

OT: American Politics

I just don't understand enough about how polling is calculated and how sophisticated they are in their modelling. How do they correct for differences in response rates by age, gender, race, socio economic?

This election will be a real bellwether for polling usefulness and if I had to guess I think the worlds going to realize how little attention we should give to it in the future.
 
I just don't understand enough about how polling is calculated and how sophisticated they are in their modelling. How do they correct for differences in response rates by age, gender, race, socio economic?

This election will be a real bellwether for polling usefulness and if I had to guess I think the worlds going to realize how little attention we should give to it in the future.

Me neither, but I want to start a polling company like Trump described.

Charge a few million, and hire a PowerPoint nerd to come up with 78 pages of fancy graphs showing how one side is leading 49 to 47
 
Me neither, but I want to start a polling company like Trump described.

Charge a few million, and hire a PowerPoint nerd to come up with 78 pages of fancy graphs showing how one side is leading 49 to 47

We really should start a fake polling firm.
 
I just don't understand enough about how polling is calculated and how sophisticated they are in their modelling. How do they correct for differences in response rates by age, gender, race, socio economic?

This election will be a real bellwether for polling usefulness and if I had to guess I think the worlds going to realize how little attention we should give to it in the future.

They try to make it super dense and complicated sounding, and sadly there is math involved, but it's basically them taking what they believe the composition of the electorate is going to be, and adjusting their survey/poll results accordingly.

So if your math is predicting that women will be 51% of the electorate and 55% of your respondents were women, you're weighting their impact on your survey downwards until they only make up 51% of the votes in your poll. Same with race, political affiliation, age, etc.
 
I love how different people's worlds are. MAGA and lesser hardcore conservatives are celebrating how amazing trump looked and libs are celebrating how bad trump looked. All based on a few 30 second clips of a 3 hour interview. I'll assume the truth is somewhere in between.
 
Me neither, but I want to start a polling company like Trump described.

Charge a few million, and hire a PowerPoint nerd to come up with 78 pages of fancy graphs showing how one side is leading 49 to 47
03trafalgar-promo-facebookJumbo-v2.jpg

3pIdWG2.png
 
I love how different people's worlds are. MAGA and lesser hardcore conservatives are celebrating how amazing trump looked and libs are celebrating how bad trump looked. All based on a few 30 second clips of a 3 hour interview. I'll assume the truth is somewhere in between.

I assume Joe didn’t call Trump out on his Hitler and “enemy within” rhetoric but happy he asked see evidence of election fraud

I also think Trump sitting down for 3 hours without yelling humanizes him and makes him seem less scary… so that’s a win for him
 
I just don't understand enough about how polling is calculated and how sophisticated they are in their modelling. How do they correct for differences in response rates by age, gender, race, socio economic?

This election will be a real bellwether for polling usefulness and if I had to guess I think the worlds going to realize how little attention we should give to it in the future.
from what I can gather, seems like a lot of them have baked in the 'hidden trump voter' into their assumptions. they existed in 2016 and 2020, dunno if they still exist post-Dobbs.
 
Why? What about MAGA's view of Trump leads you to believe that there's any rationality in their takes on his appearances?

This is extra strength fentanyl levels of bothsidesism
Trump is Trump. We simply don't have the best judgement of understanding his appeal. If Republicans (not maga Republicans) think he did well and we, who wouldn't vote for trump anyway, think he did poorly, then I'm not sure we're the best judges in this case.
 
from what I can gather, seems like a lot of them have baked in the 'hidden trump voter' into their assumptions. they existed in 2016 and 2020, dunno if they still exist post-Dobbs.

2016 definitely missed Trump voters for 2-3%. So the polls were all within MoE, but they all broke wrong in the same direction. 2020 though was such a fucked up election, I don't know if there are any polling lessons from it. Trump overperformed crazy in a few places, underperformed crazy in some others but the election was smack in the middle of a historic pandemic that caused significant changes in how people voted (mail in vs in person).
 
Back
Top