2020 vs 2024 poll averages

[Slavister]

Remember when Joe Biden barely won in 2020 by just 44,000 votes in Georgia, Arizona and Wisconsin?
This was the polling (which traditionally overestimates Democrats). Around this time in July, Biden led by ~9.6 points.



source

This is now


source


This was not just national. Click on any swing state on both sites and you get a similar kind of picture.

E.g. at those 3 closest states as of July 10,

AZ 3.0 lead by Biden in 2020 vs 4.5 lead by Trump in 2024
GA 1.7 lead by Biden in 2020 vs 5.3 lead by Trump in 2024
WI 8.6 lead by Biden in 2020 vs 1.4 lead by Trump in 2024

Only problem is - other Dem candidates do not appear to be doing much better vs Trump.

All available indications IMO is that Traitor Trump will be our next President. Hoping for some miracle in the next 4 months...

There will be those who will say - see Biden won by much less than polls predicted and therefore polls are garbage anyway. I would say that (a) Dems tend to do better in polls than Reps (for RV polls), so this argument is somewhat in line with 2020; and (b) sure, we can hope that polls are way off in the other direction but they still look very bleak for Dems compared to 2020.

 

[PoliSciXtina]

Just ignore the damn polls. They were wrong in 2016 too. and 2022.

 

[samiam5211]

Other candidates would be able to motivate voters to turn out.

Biden isn't going to motivate anyone.

Anyone motivated by fear of Trump is already baked in to the numbers.

You can send Obama/Newsome/Harris out to stump, but that's not really going to get many to go out and vote.

People want change, but they don't have the option to choose change.

 

[TangoNoCash]

With slim margins and the electoral college I am a man of little faith as it pertains to current poll predictions. It’s gonna be a stressful fall season.

 

[reflechissez]

Other candidates would be able to motivate voters to turn out.

Biden isn't going to motivate anyone.

Anyone motivated by fear of Trump is already baked in to the numbers.

You can send Obama/Newsome/Harris out to stump, but that's not really going to get many to go out and vote.

People want change, but they don't have the option to choose change.

How do you know any of this?

 

[samiam5211]

How do you know any of this?

What a stupid question.

 

[Hello My Son]

Just ignore the damn polls. They were wrong in 2016 too. and 2022.

Why would they have been so wrong in Hillary's favor in 2016, Biden's favor in 2020, but suddenly extremely wrong in Trump's favor in 2024? That doesn't pass the basic sniff test as far as logic goes.

 

[Finch]

Remember when Joe Biden barely won in 2020 by just 44,000 votes in Georgia, Arizona and Wisconsin?
This was the polling (which traditionally overestimates Democrats). Around this time in July, Biden led by ~9.6 points.

View attachment 67519471

source

This is now
View attachment 67519473

source


This was not just national. Click on any swing state on both sites and you get a similar kind of picture.

E.g. at those 3 closest states as of July 10,

AZ 3.0 lead by Biden in 2020 vs 4.5 lead by Trump in 2024
GA 1.7 lead by Biden in 2020 vs 5.3 lead by Trump in 2024
WI 8.6 lead by Biden in 2020 vs 1.4 lead by Trump in 2024

Only problem is - other Dem candidates do not appear to be doing much better vs Trump.

All available indications IMO is that Traitor Trump will be our next President. Hoping for some miracle in the next 4 months...

There will be those who will say - see Biden won by much less than polls predicted and therefore polls are garbage anyway. I would say that (a) Dems tend to do better in polls than Reps (for RV polls), so this argument is somewhat in line with 2020; and (b) sure, we can hope that polls are way off in the other direction but they still look very bleak for Dems compared to 2020.

"Political junkies get into polls like it's a sports game or something."

Polling is notoriously unreliable, they can be interesting, just not reliable especially if the race is within 6 points, let alone 3. A candidate needs to be ahead at least 6% points or more for a poll to be confident that the lead is not simply the result of a sampling error.

If it were anyone other than Trump running I would be worried for Democrats, but Trump himself will bring out a lot of blue votes regardless of who he is running against. I keep saying this, but I don't understand the logic of Trump and his Maga crowd and how cocky they are right now. I believe they will be surprised again in November.

 

[Cork]

Remember when Joe Biden barely won in 2020 by just 44,000 votes in Georgia, Arizona and Wisconsin?
This was the polling (which traditionally overestimates Democrats). Around this time in July, Biden led by ~9.6 points.

View attachment 67519471

source

This is now
View attachment 67519473

source


This was not just national. Click on any swing state on both sites and you get a similar kind of picture.

E.g. at those 3 closest states as of July 10,

AZ 3.0 lead by Biden in 2020 vs 4.5 lead by Trump in 2024
GA 1.7 lead by Biden in 2020 vs 5.3 lead by Trump in 2024
WI 8.6 lead by Biden in 2020 vs 1.4 lead by Trump in 2024

Only problem is - other Dem candidates do not appear to be doing much better vs Trump.

All available indications IMO is that Traitor Trump will be our next President. Hoping for some miracle in the next 4 months...

There will be those who will say - see Biden won by much less than polls predicted and therefore polls are garbage anyway. I would say that (a) Dems tend to do better in polls than Reps (for RV polls), so this argument is somewhat in line with 2020; and (b) sure, we can hope that polls are way off in the other direction but they still look very bleak for Dems compared to 2020.

ABC News – Breaking News, Latest News and Videos

ABC News is your trusted source on political news stories and videos. Get the latest coverage and analysis on everything from the Trump presidency, Senate, House and Supreme Court.

projects.fivethirtyeight.com

2022 senate projections.

 

[BirdinHand]

Why would they have been so wrong in Hillary's favor in 2016,

They predicted Hillary winning - she lost.

Biden's favor in 2020,

They predicted a larger margin of victory - they were wrong.

but suddenly extremely wrong in Trump's favor in 2024? That doesn't pass the basic sniff test as far as logic goes

Polls haven’t been reliable in a long time because poll methods are unreliable.

 

[Hello My Son]

They predicted Hillary winning - she lost.

Yes, that's exactly what I said. A poll that overestimates one's actual performance means that the poll was biased in that person's favor. Polling results that underestimate one's actual performance means that the poll is demonstrating bias against that person.

They predicted a larger margin of victory - they were wrong.

Yes, that means the polling estimates were demonstrating a bias in Biden's favor because the polls incorrectly OVERESTIMATED his actual performance. When polling results predicted that Hillary and Biden would perform better than their actual performances, it meant that the poll was giving them an inaccurate advantage. Otherwise known as being biased in their favor.

Polls haven’t been reliable in a long time because poll methods are unreliable.

Again, so if the premise is that polls were incorrectly biased in a way that padded Hillary's numbers in 2016 and Biden's numbers in 2020, what has changed in the methodology that would explain their bias to completely reverse course in 2024 and pad Trump's numbers instead of Biden's? That just doesn't make any sense unless they've completely changed their methodologies in ways that would cause polling to go from overestimating Biden's performance in 2020 to underestimating it in 2024, while going from underestimating Trump's performance in BOTH 2016 AND 2020, but somehow overestimating it in 2024.

What are they doing so differently to create this complete shift in overall bias, and why is this bias being demonstrated across the board, from all pollsters, in Trump's favor?

 

[BirdinHand]

Yes, that's exactly what I said. A poll that overestimates one's actual performance means that the poll was biased in that person's favor. Polling results that underestimate one's actual performance means that the poll is demonstrating bias against that person.



Yes, that means the polling estimates were demonstrating a bias in Biden's favor because the polls incorrectly OVERESTIMATED his actual performance. When polling results predicted that Hillary and Biden would perform better than their actual performances, it meant that the poll was giving them an inaccurate advantage. Otherwise known as being biased in their favor.



Again, so if the premise is that polls were incorrectly biased in a way that padded Hillary's numbers in 2016 and Biden's numbers in 2020, what has changed in the methodology that would explain their bias to completely reverse course in 2024 and pad Trump's numbers instead of Biden's? That just doesn't make any sense unless they've completely changed their methodologies in ways that would cause polling to go from overestimating Biden's performance in 2020 to underestimating it in 2024, while going from underestimating Trump's performance in BOTH 2016 AND 2020, but somehow overestimating it in 2024.

What are they doing so differently to create this complete shift in overall bias, and why is this bias being demonstrated across the board, from all pollsters, in Trump's favor?

You’re misunderstanding completely - polls are not accurately reflecting voting. For quite some time now.

Because they aren’t capturing a truly representative sampling of voters. They’re reaching the people that want to take part in polls.

How many polls have you received or answered?

 

[Hello My Son]

You’re misunderstanding completely - polls are not accurately reflecting voting. For quite some time now.

Because they aren’t capturing a truly representative sampling of voters. They’re reaching the people that want to take part in polls.

How many polls have you received or answered?

No, I understand completely. Are you suggesting that polling methodologies have changed since 2016 in ways that went from inaccurately UNDERESTIMATING Trump's support in 2016/2020 to OVERESTIMATING it now, or are you suggesting that Trump voters have shifted from NOT WANTING TO take part in polls in 2016 and 2020 to WANTING TO take part in polls in 2024, while Hillary/Biden supporters went from WANTING to take part in polls in 2016/2020 to NOT WANTING to take part in polls in 2024?

Because if we're assuming that they're inaccurate now just like they were in 2016/2020, then they should be overestimating Biden's eventual performance just like they did in 2020 and Hillary's in 2016. And if they're making the same errors now that they were in 2016/2020, then Biden is really in the toilet right now.

Or are you suggesting that Trump's polling is being overestimated right now?

​

 

[BirdinHand]

No, I understand completely. Are you suggesting that polling methodologies have changed since 2016 in ways that went from inaccurately UNDERESTIMATING Trump's support in 2016/2020 to OVERESTIMATING it now, or are you suggesting that Trump voters have shifted from NOT WANTING TO take part in polls in 2016 and 2020 to WANTING TO take part in polls in 2024, while Hillary/Biden supporters went from WANTING to take part in polls in 2016/2020 to NOT WANTING to take part in polls in 2024

I’m saying that polls have BEEN WRONG for years.

Period.

You’re the one trying to bring over/under estimating into the equation.

Wrong is wrong is wrong.

 

[reflechissez]

What a stupid question.

You gave unsubstantiated opinions. Next time, do better.

 

[multivita-man]

They predicted Hillary winning - she lost.

They predicted a larger margin of victory - they were wrong.

What else might this tell you?

 

[samiam5211]

You gave unsubstantiated opinions. Next time, do better.

I won’t.

I also never pretended it was anything but opinion.

 

[reflechissez]

I won’t.

I also never pretended it was anything but opinion.

Why so prickly when challenged? Everyone is not cut out for political boards.

 

[Mithrae]

Just ignore the damn polls. They were wrong in 2016 too. and 2022.

This reminds me quite strongly of climate science deniers. "We should just ignore all the best indicators we have available because they aren't perfect!" I guess maybe that would make sense if one supposed that this is just another election and it really doesn't matter too much how it goes, but I'd suggest the stakes are too high for that. Whatever the reasons may be, Biden just isn't doing good enough; he hasn't been for a long time, hasn't shown any improvement and there's nothing on the horizon to suggest there will be any improvement (for a while there might have been some hope that Trump's guilty verdict would swing things, but evidently not). Harris might not do good enough either, we don't really know - though there's every reason to expect her popularity to grow once a) she becomes the candidate to rally around rather than the VP choice to critique and b) Biden 'does the right thing' and the age issue can be fully turned around against Trump - but what we do know is that Biden isn't.

 

[BirdinHand]

What else might this tell you?

That polls are worthless.

 

[multivita-man]

That polls are worthless.

Or that if anyone is going to out-perform the polls, it's probably not going to be Biden.

They're not worthless. Campaigns use them. Spend a shit ton of money on putting them together and make important decisions based on them. They can't always predict election outcomes with precision but that doesn't mean they're worthless. They can gauge candidate strength. They're the only way to do it, really.

 

[triharder]

However this turns out, live your life to the best of your ability. What else could you possibly do other than that?

 

[samiam5211]

Why so prickly when challenged? Everyone is not cut out for political boards.

Why do you think my reply was prickly?

 

[Hello My Son]

I’m saying that polls have BEEN WRONG for years.

Period.

You’re the one trying to bring over/under estimating into the equation.

Yes, I'm bringing up relevant variables based on my advanced understanding of how statistical data analysis works.

Wrong is wrong is wrong.

No, there are different kinds of wrong. Methodologies that overestimate or over-sample for a candidate at one election do not suddenly start underestimating and/or under-sampling for the same candidate the next election. Not unless the methodologies change.

You're claiming that the polls are inaccurate and that both of these are true:
A) Nothing has changed about the polling methodologies to correct the previous inaccuracies, AND
B) That in spite of nothing changing about polling methodologies, the polling inaccuracies have completely inverted in terms of the biases of sampling and/or results.

That doesn't even begin to add up logically. If they didn't change erroneous methodologies, then the inaccuracies should still be reflecting the same biases they did in 2016 and 2020, but if that were true, then Trump's current polling would actually be an UNDERestimate of his eventual performance, just like the erroneous polling demonstrated in 2016/2020.

And that completely leaves out the fact that you premise is entirely wrong and methodologies HAVE been updated to get rid of landline reliance and rely more upon cell phones and online data collection.

 

[BirdinHand]

Yes, I'm bringing up relevant variables based on my advanced understanding of how statistical data analysis works

And that’s based on what? Specifically.

Nothing has changed about the polling methodologies to correct the previous inaccuracies, AND
B) That in spite of nothing changing about polling methodologies, the polling inaccuracies have completely inverted in terms of the biases of sampling and/or results

Correct. The methodologies are flawed because the populace has changed. And continues to change.

Using old methodologies on a changing populace has been a failure.

That doesn't even begin to add up logically. If they didn't change erroneous methodologies, then the inaccuracies should still be reflecting the same biases they did in 2016 and 2020, but if that were true, then Trump's current polling would actually be an UNDERestimate of his eventual performance, just like the erroneous polling demonstrated in 2016/2020

Sure it does. You’re operating under the premise that the methodology is still correct - the outcomes have been a failure so your assumption the methodology was correct is also false.