You are right, they might be spot on about this, although the recall vote always raises question. Partisanship is a bit different, because you have independents you have to push. The recall vote, if accurate, raises some questions for me. I have 0 evidence for this but I would be astounded if the 2026 Ohio electorate voted for Trump by 9 pts.
It might also be very early to accurately guess how strong Democratic turnout will be.
While I agree that the order of states could be off, the NYT polls are still suggesting shifts of anywhere from 8-14% from 2024 (excluding Maine, given the extenuating circumstances). I'm skeptical Dems do much better than that, and their modeled electorate shift of D+4 would be historically consistent (https://plusminus4.substack.com/p/blue-wave-polls)
It is especially the order of states that raises some questions for me. And some feeling in my stomach that the Ohio recall-vote won’t possibly be within two points of 2024.
Recall vote, is of course, a bit sketchy itself. The NYT never used to weight by recall vote, I’d be interested in why they went with that change.
Texas is unique in that it swung unusually hard for Trump in 2024. If you want a better baseline for Texas's partisanship relative to the national environment, use the 2024 Senate race and the 2022 Governor, Lieutenant Governor, and Attorney General races.
The actual national environment, as per polling, is currently D+6-8, which would mean an electorate shift of D+8-10. An electorate shift of D+4 as modeled here would mean a D+2 national environment, which would be very surprising unless one weighted to recalled vote or things change drastically in favor of Republicans between now and the election.
Texas also saw some steep turnout drops for Democrats in crucial metro areas. A trend that might reverse in 2026.
That together with hispanic reversal and Paxton's complete lack of cross-over appeal might flip Texas. Although we will have to wait and see how much of that is due to the electorate and how much of it is due to persuasion.
The easiest way to tell how much of that is electorate and how much is persuasion is to use Abbott's margin as a baseline. If Abbott loses by 3 points while Paxton also loses by 3 points, it's electorate. If Abbott wins by 10 points while Paxton loses by 15 points, it's persuasion.
A wild card is the other statewide races in Texas: Lieutenant Governor and Attorney General. The Attorney General race is back to being between politicians with negligible name recognition because Paxton is not running in it, while the Lieutenant Governor has incumbent Dan Patrick (basically Paxton minus all his scandals) facing off against a progressive Democrat with low name recognition. If the margins for these two are closer to the Gubernatorial race than the Senate race, persuasion in the Senate race was more about Paxton's scandals. If the margins for these two are closer to the Senate race, persuasion was more about disapproval of Trump and MAGA.
Now obviously, this works under the assumption that Abbott is still seen as popular and from the pre-Trump era. There is a chance that he is now seen as a MAGA Republican like Patrick and Paxton. In that case we'll need exit poll data to find out whether it was electorate or persuasion.
If I'm understanding you correctly, why assume the 8-10% shift is entirely driven by the electorate when this hasn't been true of past midterms (or of current polling)? Registered voter polls right now are still showing D+4 environments. 2010 was an R+15 swing from 2008—the likely voter effect was about R+7, and the rest was persuasion.
The deal is that it's next to impossible to have a functional likely voter screen. Weight 2024 too much, and you underestimate Democrats. Voter enthusiasm is hard to translate into whether someone will actually vote. Just asking people if they will vote or not seems to require a ginormous sample size considering that it's an extra variable usually not too dependent on partisanship. And if we just go off of registered voters this early in the game we miss first-time voters and voters who have to re-register because they got purged from the voting rolls, missing enthusiasm gaps.
Yes, some of the shift is driven by persuasion, but this article shows that the electorate is off as well.
You are right, they might be spot on about this, although the recall vote always raises question. Partisanship is a bit different, because you have independents you have to push. The recall vote, if accurate, raises some questions for me. I have 0 evidence for this but I would be astounded if the 2026 Ohio electorate voted for Trump by 9 pts.
It might also be very early to accurately guess how strong Democratic turnout will be.
While I agree that the order of states could be off, the NYT polls are still suggesting shifts of anywhere from 8-14% from 2024 (excluding Maine, given the extenuating circumstances). I'm skeptical Dems do much better than that, and their modeled electorate shift of D+4 would be historically consistent (https://plusminus4.substack.com/p/blue-wave-polls)
It is especially the order of states that raises some questions for me. And some feeling in my stomach that the Ohio recall-vote won’t possibly be within two points of 2024.
Recall vote, is of course, a bit sketchy itself. The NYT never used to weight by recall vote, I’d be interested in why they went with that change.
Texas is unique in that it swung unusually hard for Trump in 2024. If you want a better baseline for Texas's partisanship relative to the national environment, use the 2024 Senate race and the 2022 Governor, Lieutenant Governor, and Attorney General races.
The actual national environment, as per polling, is currently D+6-8, which would mean an electorate shift of D+8-10. An electorate shift of D+4 as modeled here would mean a D+2 national environment, which would be very surprising unless one weighted to recalled vote or things change drastically in favor of Republicans between now and the election.
Texas also saw some steep turnout drops for Democrats in crucial metro areas. A trend that might reverse in 2026.
That together with hispanic reversal and Paxton's complete lack of cross-over appeal might flip Texas. Although we will have to wait and see how much of that is due to the electorate and how much of it is due to persuasion.
The easiest way to tell how much of that is electorate and how much is persuasion is to use Abbott's margin as a baseline. If Abbott loses by 3 points while Paxton also loses by 3 points, it's electorate. If Abbott wins by 10 points while Paxton loses by 15 points, it's persuasion.
A wild card is the other statewide races in Texas: Lieutenant Governor and Attorney General. The Attorney General race is back to being between politicians with negligible name recognition because Paxton is not running in it, while the Lieutenant Governor has incumbent Dan Patrick (basically Paxton minus all his scandals) facing off against a progressive Democrat with low name recognition. If the margins for these two are closer to the Gubernatorial race than the Senate race, persuasion in the Senate race was more about Paxton's scandals. If the margins for these two are closer to the Senate race, persuasion was more about disapproval of Trump and MAGA.
Now obviously, this works under the assumption that Abbott is still seen as popular and from the pre-Trump era. There is a chance that he is now seen as a MAGA Republican like Patrick and Paxton. In that case we'll need exit poll data to find out whether it was electorate or persuasion.
If I'm understanding you correctly, why assume the 8-10% shift is entirely driven by the electorate when this hasn't been true of past midterms (or of current polling)? Registered voter polls right now are still showing D+4 environments. 2010 was an R+15 swing from 2008—the likely voter effect was about R+7, and the rest was persuasion.
The deal is that it's next to impossible to have a functional likely voter screen. Weight 2024 too much, and you underestimate Democrats. Voter enthusiasm is hard to translate into whether someone will actually vote. Just asking people if they will vote or not seems to require a ginormous sample size considering that it's an extra variable usually not too dependent on partisanship. And if we just go off of registered voters this early in the game we miss first-time voters and voters who have to re-register because they got purged from the voting rolls, missing enthusiasm gaps.
Yes, some of the shift is driven by persuasion, but this article shows that the electorate is off as well.