KFF: Trump +4

[Redban]

https://files.kff.org/attachment/Topline-KFF-Health-Tracking-Poll-February-2024.pdf

Biden +4 in the 2020 recall vote, which decreases the likelihood of non-response bias (sorry poll un-skewers)

Trump 41
Biden 37
Other 14
Won’t vote 7

[Birdish]

https://files.kff.org/attachment/Topline-KFF-Health-Tracking-Poll-February-2024.pdf

Biden +4 in the 2020 recall vote, which decreases the likelihood of non-response bias (sorry poll un-skewers)

Trump 41
Biden 37
Other 14
Won’t vote 7

You can keep repeating that over and over, it doesnt make it true.

[Redban]

https://files.kff.org/attachment/Topline-KFF-Health-Tracking-Poll-February-2024.pdf

Biden +4 in the 2020 recall vote, which decreases the likelihood of non-response bias (sorry poll un-skewers)

Trump 41
Biden 37
Other 14
Won’t vote 7

You can keep repeating that over and over, it doesnt make it true.

You can keep claiming that “oh, people are just lying and saying they voted Biden in 2020,” and it won’t enable you to  un-skew the polls any better

The whole foundation of polling is that most people will tell the truth; if people will lie, pollsters need to change methods of polling for accuracy. Not everyone will tell the truth, but to the extent lies will be in poll results, those lies shouldn’t be statistically significant. Once you go into the depths of “what if people are lying about their answers,” then everything in the entire poll becomes suspect (not just the answers you want to dismiss)

[wbrocks67]

Honestly 2020 recall matters literally nothing when you have 21% saying other or that they won't vote. That's a very useless result.

[jaichind]

KFF mostly focuses and polls on healthcare issues

https://twitter.com/Politics_Polls/status/1760855984480645380
https://twitter.com/Politics_Polls/status/1760856310726230089

[kwabbit]

https://files.kff.org/attachment/Topline-KFF-Health-Tracking-Poll-February-2024.pdf

Biden +4 in the 2020 recall vote, which decreases the likelihood of non-response bias (sorry poll un-skewers)

Trump 41
Biden 37
Other 14
Won’t vote 7

You can keep repeating that over and over, it doesnt make it true.

It's mostly true. Weighting on 2020 recall is one of the better ways to eliminate non-response bias. You're at least getting the right number of Biden/Trump 2020 voters. The issue would be if Biden 2020, Trump 2024 voters are more likely to answer surveys, which is possible given that the best explanation of the 2020 polling error is that Trump 2016, Biden 2020 voters answered the polls at unusually high rates. However the Democratic base is increasingly educated and has higher social trust than the Republican base, making them more likely to answer polls, decreasing the chance of non-response bias. A more likely source of polling error would be two-fold: an underestimate of Biden's numbers with non-Whites and a turnout advantage in part from GOP early voting fears. Very plausible but still a hunch motivated by partisanship.

[The Economy is Getting Worse]

https://files.kff.org/attachment/Topline-KFF-Health-Tracking-Poll-February-2024.pdf

Biden +4 in the 2020 recall vote, which decreases the likelihood of non-response bias (sorry poll un-skewers)

Trump 41
Biden 37
Other 14
Won’t vote 7

You can keep repeating that over and over, it doesnt make it true.

It's mostly true. Weighting on 2020 recall is one of the better ways to eliminate non-response bias. You're at least getting the right number of Biden/Trump 2020 voters. The issue would be if Biden 2020, Trump 2024 voters are more likely to answer surveys, which is possible given that the best explanation of the 2020 polling error is that Trump 2016, Biden 2020 voters answered the polls at unusually high rates. However the Democratic base is increasingly educated and has higher social trust than the Republican base, making them more likely to answer polls, decreasing the chance of non-response bias. A more likely source of polling error would be two-fold: an underestimate of Biden's numbers with non-Whites and a turnout advantage in part from GOP early voting fears. Very plausible but still a hunch motivated by partisanship.

Also it would have to involved Trump - Downballot D voters answering at unusually high rates somehow. The fundamentals are just very bad for Biden, with a worsening economy, increasing foreign conflicts, and a migrant crisis. Biden has just done a very bad job.

[kwabbit]

https://files.kff.org/attachment/Topline-KFF-Health-Tracking-Poll-February-2024.pdf

Biden +4 in the 2020 recall vote, which decreases the likelihood of non-response bias (sorry poll un-skewers)

Trump 41
Biden 37
Other 14
Won’t vote 7

You can keep repeating that over and over, it doesnt make it true.

It's mostly true. Weighting on 2020 recall is one of the better ways to eliminate non-response bias. You're at least getting the right number of Biden/Trump 2020 voters. The issue would be if Biden 2020, Trump 2024 voters are more likely to answer surveys, which is possible given that the best explanation of the 2020 polling error is that Trump 2016, Biden 2020 voters answered the polls at unusually high rates. However the Democratic base is increasingly educated and has higher social trust than the Republican base, making them more likely to answer polls, decreasing the chance of non-response bias. A more likely source of polling error would be two-fold: an underestimate of Biden's numbers with non-Whites and a turnout advantage in part from GOP early voting fears. Very plausible but still a hunch motivated by partisanship.

Also it would have to involved Trump - Downballot D voters answering at unusually high rates somehow. The fundamentals are just very bad for Biden, with a worsening economy, increasing foreign conflicts, and a migrant crisis. Biden has just done a very bad job.

Good point with the Senate polling. It's hard to say that a Gallego +10 poll is not capturing enough Democrats.

Sometimes non-response bias is a valid theory but in many cases it's just "This poll has Biden down 4 and I think he's actually up 2 so it's not capturing enough Biden voters to satisfy my partisan inclinations".