2024 Political Polls

  • Thread starter Thread starter nycfan
  • Start date Start date
  • Replies: 2K
  • Views: 55K
  • Politics 
05C40 said:
Which raises the question, how are polls conducted these days? I'm 70 and don't have a landlines, don't answer unknown numbers, and delete without reading any emails/texts I don't recognize. And I am in the age group that is supposedly oversampled?
Click to expand...
I'm a couple of years older and do exactly the same thing.
 
lawtig02 said:
I’m sorry but this is dumb as hell in light of recent high quality polling.
Click to expand...
This reminds me of the folks saying that Nate was dumb as hell giving Trump a 25% chance in 2016.

Remember: Biden won the popular vote by 4.5% and barely won the election in the EC. I know you know that, because above you wrote that the lead needs to be 5-7 to feel comfortable headed into election day.

I feel pretty confident that Nate's model is solid. It isn't necessarily better than the current 538 model, but there's no reason to think that Nate made his model worse. You can question his choices about the inclusion of junk polling, but there are good arguments both ways on that.
 
superrific said:
Nate Silver has his issues, but I see no reason to think that he's putting his thumb on the scale in his model. It would be professional suicide.
Click to expand...
I don't think he is.

I just find him repulsive at this point. Just in it for the $$ and the fame. Dude loves to get clicks and attention. He didn't used to be like that. But it got exponentially worse after ABC let him go
 
superrific said:
This reminds me of the folks saying that Nate was dumb as hell giving Trump a 25% chance in 2016.

Remember: Biden won the popular vote by 4.5% and barely won the election in the EC. I know you know that, because above you wrote that the lead needs to be 5-7 to feel comfortable headed into election day.

I feel pretty confident that Nate's model is solid. It isn't necessarily better than the current 538 model, but there's no reason to think that Nate made his model worse. You can question his choices about the inclusion of junk polling, but there are good arguments both ways on that.
Click to expand...
Let’s just step back from the polling for a second. Nate’s model just gave Trump his highest odds of winning at any point in this campaign — EVER. Higher than after the catastrophic Biden debate. Higher than after Trump was almost assassinated. Higher than anytime in the following weeks when Biden was a dead man walking. Higher than at any point before Kamala’s recent surge.

What could possibly have happened in the last couple of weeks that would move the election so strongly in Trump’s favor? There’s literally nothing. If Kamala bombs tomorrow’s debate, that could do it. If Russia nukes Kiev, that could do it. If Covid 4.0 — The Reckoning, starts laying waste to our cities over the next couple of months, that could do it. But there’s just no reason a model should be swinging that much based on anything that has actually happened in real life.

I don’t have any beef against Silver. I agree he’s trying to do a good job. But whatever he’s picking up on now is almost certainly BS, and I don’t see any way around that.
 
aGDevil2k said:
I don't think he is.

I just find him repulsive at this point. Just in it for the $$ and the fame. Dude loves to get clicks and attention. He didn't used to be like that. But it got exponentially worse after ABC let him go
Click to expand...
1. In fairness, you kind of have to do that if you're trying to make a living off substack. When he built 538, he was the only game in town. No longer.
2. OTOH I think he considers himself a professional poker player at this point. No idea how much he makes off that. So maybe the professional suicide isn't as big a deal to him.
 
lawtig02 said:
Let’s just step back from the polling for a second. Nate’s model just gave Trump his highest odds of winning at any point in this campaign — EVER. Higher than after the catastrophic Biden debate. Higher than after Trump was almost assassinated. Higher than anytime in the following weeks when Biden was a dead man walking. Higher than at any point before Kamala’s recent surge.

What could possibly have happened in the last couple of weeks that would move the election so strongly in Trump’s favor? There’s literally nothing. If Kamala bombs tomorrow’s debate, that could do it. If Russia nukes Kiev, that could do it. If Covid 4.0 — The Reckoning, starts laying waste to our cities over the next couple of months, that could do it. But there’s just no reason a model should be swinging that much based on anything that has actually happened in real life.

I don’t have any beef against Silver. I agree he’s trying to do a good job. But whatever he’s picking up on now is almost certainly BS, and I don’t see any way around that.
Click to expand...
1. There is still a convention bounce in there for Kamala, as I understand it. It's small, but it's present. So maybe Trump would be doing 5% worse if not for that.
2. More importantly, the model is designed to do exactly this. I don't think I need to go through a hypo for you, but I'll do so for the sake of the general reading population.

Suppose there's a presidential race, and one candidate is winning by 10 points in February. What are the odds that the candidate will win the election? If you go by the polls alone, probably close to 100%. It would take a massive polling error to make up 10 points, even with the EC. But it's still Feb, which means a lot can change. I would guess that the model would give about a 70% chance for the winning candidate to win -- the other 30% being uncertainty about what will happen in the future.

Now, over time, the polls do narrow a bit, but on November 1, the leading candidate is still up 6. The odds of that candidate winning the election would be much higher, like 90%+. The reduced uncertainty more than compensates for the reduced poll margin. At this point, really the only thing the losing candidate can hope for is a really pronounced EC effect or a large polling error.

3. In a way, this is a lot like pricing stock options. A 1-Y call option with a strike price of 50 for a stock trading at 40 will usually trade for more than a 1-mo option with a strike price of 45, even though it's less favorable on its own terms. The time to maturity makes a huge difference.
 
superrific said:
1. There is still a convention bounce in there for Kamala, as I understand it. It's small, but it's present. So maybe Trump would be doing 5% worse if not for that.
2. More importantly, the model is designed to do exactly this. I don't think I need to go through a hypo for you, but I'll do so for the sake of the general reading population.

Suppose there's a presidential race, and one candidate is winning by 10 points in February. What are the odds that the candidate will win the election? If you go by the polls alone, probably close to 100%. It would take a massive polling error to make up 10 points, even with the EC. But it's still Feb, which means a lot can change. I would guess that the model would give about a 70% chance for the winning candidate to win -- the other 30% being uncertainty about what will happen in the future.

Now, over time, the polls do narrow a bit, but on November 1, the leading candidate is still up 6. The odds of that candidate winning the election would be much higher, like 90%+. The reduced uncertainty more than compensates for the reduced poll margin. At this point, really the only thing the losing candidate can hope for is a really pronounced EC effect or a large polling error.

3. In a way, this is a lot like pricing stock options. A 1-Y call option with a strike price of 50 for a stock trading at 40 will usually trade for more than a 1-mo option with a strike price of 45, even though it's less favorable on its own terms. The time to maturity makes a huge difference.
Click to expand...
Yes, I agree with that analysis of the modeling. I just don’t get what Nate’s model is seeing, especially anticipating the next two months. If he had Kamala favored 60-40, or even 55-45, I would get it. But I don’t see anything in the data, or in future prognostications, that should put Trump ahead in the modeling by his largest margin to date.
 
lawtig02 said:
Let’s just step back from the polling for a second. Nate’s model just gave Trump his highest odds of winning at any point in this campaign — EVER. Higher than after the catastrophic Biden debate. Higher than after Trump was almost assassinated. Higher than anytime in the following weeks when Biden was a dead man walking. Higher than at any point before Kamala’s recent surge.

What could possibly have happened in the last couple of weeks that would move the election so strongly in Trump’s favor? There’s literally nothing. If Kamala bombs tomorrow’s debate, that could do it. If Russia nukes Kiev, that could do it. If Covid 4.0 — The Reckoning, starts laying waste to our cities over the next couple of months, that could do it. But there’s just no reason a model should be swinging that much based on anything that has actually happened in real life.

I don’t have any beef against Silver. I agree he’s trying to do a good job. But whatever he’s picking up on now is almost certainly BS, and I don’t see any way around that.
Click to expand...
What Trump is doing, you're not seeing, because you are not in a media bubble primarily comprised of one of the following 4 constituencies (all of which represent "secret" upsides where trump can gain new voters outside of what we view as his hard upside limit:
  • protestant (evangelical) Hispanics
  • disaffected young black men
  • disaffected bro culture young (primarily white) men
  • Anti-communist hispanics (e.g. Cubans, but many others; here's where the Komrade Kamala attacks and memes are pointed)
He's got someone doing data for him that is very good at identifying these constituencies where he can eke out new votes (and I question how organic that is to his campaign, but that's a different story).

In any event modest gains in those constituencies could swing a close election. Take a moment to be scared, and then go work your ass off to save the country you love. If you've never knocked on doors before it may seem intimating, but if an introvert like me can do it, you can too.
 
Ddseddse said:
What Trump is doing, you're not seeing, because you are not in a media bubble primarily comprised of one of the following 4 constituencies (all of which represent "secret" upsides where trump can gain new voters outside of what we view as his hard upside limit:
  • protestant (evangelical) Hispanics
  • disaffected young black men
  • disaffected bro culture young (primarily white) men
  • Anti-communist hispanics (e.g. Cubans, but many others; here's where the Komrade Kamala attacks and memes are pointed)
He's got someone doing data for him that is very good at identifying these constituencies where he can eke out new votes (and I question how organic that is to his campaign, but that's a different story).

In any event modest gains in those constituencies could swing a close election. Take a moment to be scared, and then go work your ass off to save the country you love. If you've never knocked on doors before it may seem intimating, but if an introvert like me can do it, you can too.
Click to expand...
I hear you, and I’m doing the work. But if I were to compose a list titled “Groups Most Likely To Bitch Loudly and Even Answer Pollster Calls But Not Actually Show Up At The Polls,” it would look a lot like yours, save the Cubans.
 
CFordUNC said:
Why would selecting Shapiro as VP “guarantee” a win in PA but Shapiro campaigning even more frequently in PA on Harris’s behalf while not on the ticket, not guarantee it?

I don’t think you have to wonder at all if Kamala regrets her VP pick. Her selection of Tim Walz has been even more resoundingly spectacular of a success than even the most hopeful and optimistic Democratic partisan could have envisioned.
Click to expand...
Ok, guarantee is an exaggeration, but in a state that is so close and so important, having a very popular governor on the ticket would likely make a difference. I'm acknowledging that VP selection isn't shown to have a huge impact in limited research, at least through 2008, but the political landscape is much different now than it was then.
 
538's latest:

"Vice President Kamala Harris got a few scares in the polls over the weekend. A new national poll from The New York Times and Siena College shows Harris down two points versus Former President Donald Trump among likely voters. And surveys in Michigan, Pennsylvania and Wisconsin from YouGov and CBS News found Harris and Trump within the margin of error among likely voters. These new polls confirm a slight trend towards Trump over the last two weeks, moving our forecast of the race closer to a pure toss-up.

The big thing to watch this week will be any post-debate polls from the big pollsters. I expect a quick turnaround of national polls from online panels later in the week, national live-phone polls to come out Friday and over the weekend, and state polls over the weekend and into next week. Any significant changes in the polls could impact our forecast somewhat aggressively, since we are now in the stage of the campaign where remaining volatility is declining rapidly as the time left for polls to change vanishes.

Still, it would take quite a shift to move the forecast into the 70+ percent probability range for either candidate; given recent misses in pre-election polls, our forecast factors in a relatively high probability of a correlated, large bias among most polls. Such a miss would cause averages across states to overstate support for one candidate by similar amounts (sound familiar?). If the campaign stays roughly the way it’s been for months, expect the model to see a close race through Election Day.

There is a 54-in-100 chance that Harris wins at least one state Biden didn't win in 2020
See a new scenario
random-icon.svg

538’s forecast is based on a combination of polls and campaign “fundamentals,” such as economic conditions, state partisanship and incumbency. It’s not meant to “call” a winner, but rather to give you a sense of how likely each candidate is to win. Check out our methodology to learn exactly how we calculate these probabilities."
 
You must log in or register to reply here.
Back
Top