2024 US Presidential Election: First Polls Close in 24 Hours

And then what happened

What happened after this

Please tell us

oh, i was trying to make eye contact. i was staring straight at her when i spit. probably got lucky she was trying her hardest not to look at me.

September: Sure bud, whatever you say.
November: Holy shit.

https://twitter.com/dobdob365/status/1839417418978107855?t=z8kQktaNykIAkbc2WB4UCA&s=19

9 Likes

14 Likes
2 Likes

First of all how dare you?

Woohoo, got one

Love you guys :rofl:

1 Like

3 Likes

Iowa and Kansas could go blue. I know my people.

1 Like

You are neglecting

  1. There are a lot of legit bad actors out there putting out polling. This includes both not-explicitly-partisan pollsters as well as outright partisan operations who deliberately put out Republican leaning data. Their motivations may be mixed, whether it’s “Everyone was too blue in 2016 and 2020, so if we just put out more Trumpy numbers we’ll look like geniuses!” or deliberately trying to sabotage Dem morale and turnout or perhaps something else, but we know outfits are publishing results that are redder than what the same underlying raw data run through 2012 weighting would result in.

All the results from the bad actors is useless, and anyone herding with the bad actors, which is a lot of other outfits, is also useless.

Selzer doesn’t fuck around and doesn’t herd. When her polls in 2008 and 2012 were near the consensus, she and the consensus were right. When she was the outlier in 2016 and 2020, she was right and the consensus was the outlier. It’s a peerless (albeit hyperfocused, which does make success easier) track record of success. An unimpeachable signal amidst a ton of noise is worth a lot of weight.

11 Likes

Isn’t a peerless track record of success over 4 events probably just luck?

1 Like

Not when you throw in that whites are swinging Harris and POC swinging Trump, and Iowa is the whitest state

Of course. People are very bad at recognizing small sample size, especially when it’s over a long length of time. It’s part of why bobman insists Elon Musk has rare business skill levels

Iowa is like the 8th whitest. It is the whitest state that could be considered a swing (maybe Maine which is second whitest).

It’s definitely possible that it’s just luck. It’s also a bigger sample size than Wookie posted, and is an indication her methods aren’t pure crap - especially given she doesn’t herd like most pollsters.

Vibes reduced with final NYT / Siena poll which is either herded or bad and I don’t have the energy to decide which.

https://x.com/sienaresearch/status/1853014255588843801?s=46&t=XGja5BtSraUljl_WWUrIUg

1 Like

doesn’t seem bad?

1 Like

Ok but it’s still something I’ve seen nobody addressing: one of the reasons, if not the biggest reason, GOP doing poorly in IA is such a strong signal they’re doing poor everywhere, is because of all the white voters. So is the white voter shift getting counted twice or not, when people let this IA poll reshape everything? And the POC voter shift discounted to near zero?

Not terrible, just weird and closer than I’d like.

I wouldn’t be so quick to label them as “bad actors”. Yes those certainly do exist, but there are probably some trying their best and don’t know what to do.

Let’s say you’re a Pennsylvania polling firm and for the last two elections, you had Hillary +5 and Biden +10, when the real results were a Hillary loss and a Biden squeaker. Now you see Kamala +9. What do you do? I don’t think the answer is super obvious.

This Iowa place doesn’t have that problem. If they miss badly here, it can be dismissed as a one-off.

1 Like