The question is whether the polling averages are accurately measuring the race. Right now the PA polling average on 538 includes a bunch of right-leaning polls (Trafalgar; Insider Advantage; Patriot Polling) that give Trump a lead. These same polls led 538 to give Oz a lead of 0.5% in 2022. Even though 538 tries to correct for house effe…
The question is whether the polling averages are accurately measuring the race. Right now the PA polling average on 538 includes a bunch of right-leaning polls (Trafalgar; Insider Advantage; Patriot Polling) that give Trump a lead. These same polls led 538 to give Oz a lead of 0.5% in 2022. Even though 538 tries to correct for house effects/biases of polls, they cannot totally negate the influence of red-leaning polls on their aggregation. So if Harris is leading in PA 10, this could (but need not) mean that her lead in PA is greater than what 538 is presently showing.
Yeah I agree that the current PA polling average is skewed by a bunch of right-leaning polls. However, even if we look at only the most reputable polls, I don't think any of them suggests KH is leading in PA by 9 - 10 points. If the 9 point swing in PA-10 is even close to being real, that means there's likely slippage elsewhere is all I'm saying.
A 9-10 point swing in one district need not mean a 9-10 point swing statewide. There might be relatively few districts like PA 10: suburbanish district that had not yet swung heavily towards Democrats. Most of the Philly suburban districts have already swung towards the Dems, with Bucks county showing the weakest swing. Parts of PA 10 are somewhat like the Philly suburbs: highly educated voters who might have supported Republicans in the past. If those voters are slower to move than their Philly area counterparts, and there are few if any other districts that are late swinger, then this swing might be somewhat isolated.
Hence a 9-10 point swing in PA 10 might only translate into a 1-3 point swing statewide, even without any countervailing swing in other areas towards Trump.
Yeah, we don't really disagree. I guess what I was saying was PA-10 represents a certain type of geographic/demographic, suburb/exurb. Its significant swing towards the left needs to be offset by the rightward swing in other dems/geographic areas to keep the entire state on the same level of competitiveness as 2020.
I'm increasingly convinced that it's wishful thinking that the 2022 electorate is gonna show up, in Pennsylvania and elsewhere. Midterm turnout has always skewed upscale and highly educated. In the past, that's meant Republican-leaning, but especially after Dobbs, it meant an electorate disproportionately friendly to Democrats in the new coalition. If more downscale and less educated voters show up in 2024, I suspect it cancels out the midterm turnout model and looks more like the 2016 and 2020 electorates that were much more Republican. Sure, Biden still won with the 2020 electorate, but after four years of additional downscale, less-educated voter realignment, I'm increasingly suspicious we'll look back on swing state polling as being too generous to Trump.
I bet the D friendly upscale electorate is not constant either. Would be a realignment further to D?
You might be correct that the downscale realignment Redding is bigger than the upscale realignment bluing, in the rust belt. I am not sure it is the case in GA/NC/TX where the downscale already went thru southern realignment but the upscale realignment is ongoing.
BTW, if there is no realignment, just pure migration redistribution, then a constant country level, some states Redding doesn’t mean there is always another state bluing. Think if we had only two states CA/TX, and the net migration CA->TX being redder than CA but bluer than TX. A static country level means both CA and TX getting bluer.
Right. I should have clarified that I was talking mostly about Pennsylvania in that response. It tracks some nationally as well, but not nearly as much in Texas or certainly Georgia, where the D-friendly upscale electorate is experiencing constant growth compared to those in the Rust Belt.
I probably didn't make clear enough that my comment was primarily focused on Pennsylvania rather than nationally. But even as it applies to the nonwhite working-class, plenty of tea leaves are pointing to an education-based realignment in every racial category. We'll have to wait four weeks to find out if that materializes, but it's baked into my theory that higher Hispanic turnout in Reading and Allentown than in the midterm will at this point probably net votes for Trump.
Yes, you and I fundamentally disagree on how much nonwhite working class realignment is going on, but I'll admit some swing that direction from 2020 is possible. I'll also point out that the presidential year electorate pulls in a lot of younger voters, including college students.
The question is whether the polling averages are accurately measuring the race. Right now the PA polling average on 538 includes a bunch of right-leaning polls (Trafalgar; Insider Advantage; Patriot Polling) that give Trump a lead. These same polls led 538 to give Oz a lead of 0.5% in 2022. Even though 538 tries to correct for house effects/biases of polls, they cannot totally negate the influence of red-leaning polls on their aggregation. So if Harris is leading in PA 10, this could (but need not) mean that her lead in PA is greater than what 538 is presently showing.
Yeah I agree that the current PA polling average is skewed by a bunch of right-leaning polls. However, even if we look at only the most reputable polls, I don't think any of them suggests KH is leading in PA by 9 - 10 points. If the 9 point swing in PA-10 is even close to being real, that means there's likely slippage elsewhere is all I'm saying.
Point taken.
A 9-10 point swing in one district need not mean a 9-10 point swing statewide. There might be relatively few districts like PA 10: suburbanish district that had not yet swung heavily towards Democrats. Most of the Philly suburban districts have already swung towards the Dems, with Bucks county showing the weakest swing. Parts of PA 10 are somewhat like the Philly suburbs: highly educated voters who might have supported Republicans in the past. If those voters are slower to move than their Philly area counterparts, and there are few if any other districts that are late swinger, then this swing might be somewhat isolated.
Hence a 9-10 point swing in PA 10 might only translate into a 1-3 point swing statewide, even without any countervailing swing in other areas towards Trump.
Yeah, we don't really disagree. I guess what I was saying was PA-10 represents a certain type of geographic/demographic, suburb/exurb. Its significant swing towards the left needs to be offset by the rightward swing in other dems/geographic areas to keep the entire state on the same level of competitiveness as 2020.
I'm increasingly convinced that it's wishful thinking that the 2022 electorate is gonna show up, in Pennsylvania and elsewhere. Midterm turnout has always skewed upscale and highly educated. In the past, that's meant Republican-leaning, but especially after Dobbs, it meant an electorate disproportionately friendly to Democrats in the new coalition. If more downscale and less educated voters show up in 2024, I suspect it cancels out the midterm turnout model and looks more like the 2016 and 2020 electorates that were much more Republican. Sure, Biden still won with the 2020 electorate, but after four years of additional downscale, less-educated voter realignment, I'm increasingly suspicious we'll look back on swing state polling as being too generous to Trump.
I bet the D friendly upscale electorate is not constant either. Would be a realignment further to D?
You might be correct that the downscale realignment Redding is bigger than the upscale realignment bluing, in the rust belt. I am not sure it is the case in GA/NC/TX where the downscale already went thru southern realignment but the upscale realignment is ongoing.
BTW, if there is no realignment, just pure migration redistribution, then a constant country level, some states Redding doesn’t mean there is always another state bluing. Think if we had only two states CA/TX, and the net migration CA->TX being redder than CA but bluer than TX. A static country level means both CA and TX getting bluer.
Right. I should have clarified that I was talking mostly about Pennsylvania in that response. It tracks some nationally as well, but not nearly as much in Texas or certainly Georgia, where the D-friendly upscale electorate is experiencing constant growth compared to those in the Rust Belt.
Of course, the downscale electorate is also disproportionately non-white...
I probably didn't make clear enough that my comment was primarily focused on Pennsylvania rather than nationally. But even as it applies to the nonwhite working-class, plenty of tea leaves are pointing to an education-based realignment in every racial category. We'll have to wait four weeks to find out if that materializes, but it's baked into my theory that higher Hispanic turnout in Reading and Allentown than in the midterm will at this point probably net votes for Trump.
Yes, you and I fundamentally disagree on how much nonwhite working class realignment is going on, but I'll admit some swing that direction from 2020 is possible. I'll also point out that the presidential year electorate pulls in a lot of younger voters, including college students.