Based on my analysis of their historical ratings, Cook, Inside Elections and Sabato’s Crystal Ball are probably underestimating Democrats' House win probability by about 30 percentage points
Another problem with the lowballing of Democrats by Cook, Inside Elections, and Sabato's Crystal Ball is that it gives fuel to the election deniers who will cite those forecasts as "evidence" that Democrats cheated to win.
I'm curious if this holds back into the pre-fright-wing media era. I'm also curious if there is a directionality factor at all. There is certainly a directionality in turnout of low consistency voters, which it seems to me could be interacting with past vote choice to bias perception here.
If California votes for Congressional Districts in 2024, similarly to 2018 results, the 50+1 ratings are practically spot on, and the race raters are substantially off on the races that might be in question.
Serious question here. This relates more to the Senate than the House. I’m a psychologist, not a polling expert. In that context, I’m concerned that the polls are underestimating “coming home” effects in the deep red states that will decide the Senate. My impression is that tribal/racial/media dynamics in those states make voting for a Democrat simply unthinkable for most voters. So, those (angry) voters may be telling pollsters that they’ll vote Dem, and are probably being sincere in doing so, but in the voting booth they’ll “come home” to the GOP. Would love to hear your thoughts on this.
There is also a chance that those red state voters will vote for a Libertarian or just stay home and not vote. Much depends on people actually voting. In my red state of Indiana, we are lucky to get 40% of voters to turn out (primaries are worse). Much depends on parties' catching the attention and motivating those people who do not vote.
Right. I think that’s what all the anti-trans red meat is about, trying to motivate GOP turnout. I think the GOP is thinking similarly to me: their voters won’t vote for Dems, but if they don’t vote at all that will help the Dems.
This analysis is spot on. And actually has been obvious all year long, though the analysis goes into additional dimensions. Cook and the others have been playing catch up over and over again for months. It does a huge disservice because too many people and organizations focus on their middle of the pack, and not enough on the real competitive edge where additional funding is needed.
One other thing that Nate Silver has been criticizing these qualitative race rating guys for quite a lot is that they do this thing where everything remotely contentious is a tossup until like two days before the election when they move a ton of races towards one column or the other based on essentially "oh the election is almost here, time to herd behind the consensus." They're hardcore tailists.
This analysis makes sense, but it doesn't give any weight to the Republican gerrymandering. Doesn't that affect the number of seats Dems can expect to win?
Perhaps so, but I'm afraid that Elliott's analysis doesn't because it is substantially based on the generic polling and the historic shift away from the party in power. Doesn't the fact that the party in power has just effectively stolen a half dozen or more seats offset those factors, so that a Dem advantage which might historically been expected to produce, say, a 10-seat gain for the Democrats will produce only 4 or fewer seats?
A half dozen stolen seats doesn't really matter if Dems can very reasonably expect to win, on net, 15+ seats even with those accounted for. A secular shift to a D+7 generic ballot (and this is ignoring overperformances by good candidates, sudden demographic collapses busting gerrymanders, etc.) gets you to 230 Dem seats in the House or so. Note that the numbers we have indicate huge collapses for the GOP among Hispanic voters, which makes the Texas gerrymander much weaker than it was intended to be and verging on a dummymander.
Thanks for bringing the data to something that has been intuitively obvious for a while. Another element that is also obvious is the change latency and that is a serious problem right now given accelerating fuel prices, tariff effects and also the effects from other issues that are important to large slices of the electorate, all negatives for the GOP
Careful, Elliott: You're analysis is sophisticated but I think there may be a major flaw. In examining prediction-result patterns historically, your (small) set of elections (2010, 2014, 2018, 2022) balances three elections with a Democrat in the White House (2010, 2014, 2022) against one with a Republican (2018).
There's nothing inherently wrong with that, but it contributes strongly to the overall impression that the incumbency bias is nonpartisan -- i.e., bi-directional.
A strong case can be made, however, for regarding 2018 as highly atypical -- essentially a one-off, black swan event. Karl Rove, in a fit of pique over his displacement by Steve Bannon, took the 2018 election off. And I would submit that his talents -- i.e., his dark arts -- were sorely missed. (I'd also note, parenthetically, that the 2018 blue wave yielded a net gain of 40 House seats [and a *loss* of two Senate seats] while the 2010 red wave yielded a net gain of 63 House seats for the GOP -- an asymmetry not fully explained by organic factors.)
If 2018 had as much to do with who was navigating the GOP bus as with purported generic polling bias, then we're left with a series of elections in which pollsters missed to the left. And we can add to that the 2020 election, in which of 27 House contests rated tossup by Cook, the GOP won every single one (and also won 7 Democratic leaners while losing none of their own), and similar, if less egregious misses in 2016 and 2024.
So overall, the clear pattern in the computerized voting era is pollsters overestimating Democratic strength, making corrections, and overestimating Democratic strength yet again (pace 2018). Pollsters don't stay in business by erring repeatedly and consistently *in the same direction*! So it is understandable that Cook et al are lowballing Democratic prospects in these midterms. They have touched the hot stove -- over and over -- and are very wary of doing it yet again.
You can draw your own conclusions on why they keep missing in the same direction -- why, that is, there is a pervasive polling-to-votecount red shift, even after they tweak their sampling methodologies to account for it. Maybe there really is a "reluctant Bush/Trump responder" phenomenon that results in an oversampling of Democrats (but there is demographic stratification to handle that). Or, just spitballing here, maybe unreliable vote counts (from suppression schemes and/or vote-count manipulation) are a factor? Maybe all the red thumbs on the electoral scales have left the pollsters wondering what they keep doing wrong, when they're not doing much of anything wrong at all?
This is worth our consideration as we move toward an election in which Donald Trump has called upon his side (voters, administrators, operatives, vendors...) to "cheat like hell."
What do we know about mail-in ballots relative to in person ballots? Have there been polls of already voted mail-ins one or two weeks before election day? (screening call is have you already mailed in your ballot and if yes you're in the survey. Or the question could be already voted/definitely will vote/ likely vote. Weighting the responses generates another complexity)
As is currently being noted on the talk shows voting is starting now and those votes won't change (in any significant way) from time of mailing to election day. Are there estimates of percentage of mail in ballots returned 4 weeks, 3 weeks, 2 weeks and 1 week prior to the election? Sentiment (or events) one week before the election won't change large percentage of the mail in vote.
Dammit, large GOP money well be hawking the “ewwww” factor, the fifth grade schoolyard stuff *really* hard starting in two weeks. It’s inconceivable that it will sway things — but it’s also inconceivable that it worked in 2024.
Maybe this time that will prove to be a well that had been gone to once too often? I’m still not completely over 1980, so we’ll see.
There are $6 diesel "billboards" that have been posted all over the country. There are $7/lb ground beef signs up in grocery stores. Elon can't by a $3 diesel billboard or reprice ground beef. Money can't buy you love and has significant limitations in buying votes.
I mean, that makes total sense. Selzer STILL hasn't lived down calling Iowa for Biden in 2024. Democrats are so het up at the moment that even if the magnitude is *slightly* off in the GOP's favor, pollsters are going to deal with a lot of shit. If they undersell it there's a lot less to lose.
Now I want to know who are the ~11 "safe" candidates from Inside Elections and ~3 "safe" candidates from Sabato who lost!
Another problem with the lowballing of Democrats by Cook, Inside Elections, and Sabato's Crystal Ball is that it gives fuel to the election deniers who will cite those forecasts as "evidence" that Democrats cheated to win.
I'm curious if this holds back into the pre-fright-wing media era. I'm also curious if there is a directionality factor at all. There is certainly a directionality in turnout of low consistency voters, which it seems to me could be interacting with past vote choice to bias perception here.
I remember Cook missing the 20218 wave. And that was after they had started shifting Dem late in the cycle
If California votes for Congressional Districts in 2024, similarly to 2018 results, the 50+1 ratings are practically spot on, and the race raters are substantially off on the races that might be in question.
Serious question here. This relates more to the Senate than the House. I’m a psychologist, not a polling expert. In that context, I’m concerned that the polls are underestimating “coming home” effects in the deep red states that will decide the Senate. My impression is that tribal/racial/media dynamics in those states make voting for a Democrat simply unthinkable for most voters. So, those (angry) voters may be telling pollsters that they’ll vote Dem, and are probably being sincere in doing so, but in the voting booth they’ll “come home” to the GOP. Would love to hear your thoughts on this.
There is also a chance that those red state voters will vote for a Libertarian or just stay home and not vote. Much depends on people actually voting. In my red state of Indiana, we are lucky to get 40% of voters to turn out (primaries are worse). Much depends on parties' catching the attention and motivating those people who do not vote.
Right. I think that’s what all the anti-trans red meat is about, trying to motivate GOP turnout. I think the GOP is thinking similarly to me: their voters won’t vote for Dems, but if they don’t vote at all that will help the Dems.
This analysis is spot on. And actually has been obvious all year long, though the analysis goes into additional dimensions. Cook and the others have been playing catch up over and over again for months. It does a huge disservice because too many people and organizations focus on their middle of the pack, and not enough on the real competitive edge where additional funding is needed.
When Lakshya Jain and G. Elliot Morris agree on something you know something's up.
One other thing that Nate Silver has been criticizing these qualitative race rating guys for quite a lot is that they do this thing where everything remotely contentious is a tossup until like two days before the election when they move a ton of races towards one column or the other based on essentially "oh the election is almost here, time to herd behind the consensus." They're hardcore tailists.
This analysis makes sense, but it doesn't give any weight to the Republican gerrymandering. Doesn't that affect the number of seats Dems can expect to win?
the big Republican gerrymander project was after the 2010 census. So gerrymandering was in effect in 2018
How doesn't it? The district lines are taken into account in the ratings by cook, sabato, and inside elections
Perhaps so, but I'm afraid that Elliott's analysis doesn't because it is substantially based on the generic polling and the historic shift away from the party in power. Doesn't the fact that the party in power has just effectively stolen a half dozen or more seats offset those factors, so that a Dem advantage which might historically been expected to produce, say, a 10-seat gain for the Democrats will produce only 4 or fewer seats?
A half dozen stolen seats doesn't really matter if Dems can very reasonably expect to win, on net, 15+ seats even with those accounted for. A secular shift to a D+7 generic ballot (and this is ignoring overperformances by good candidates, sudden demographic collapses busting gerrymanders, etc.) gets you to 230 Dem seats in the House or so. Note that the numbers we have indicate huge collapses for the GOP among Hispanic voters, which makes the Texas gerrymander much weaker than it was intended to be and verging on a dummymander.
That mid-district gerrymander may backfire in some Florida districts as well.
Thanks for bringing the data to something that has been intuitively obvious for a while. Another element that is also obvious is the change latency and that is a serious problem right now given accelerating fuel prices, tariff effects and also the effects from other issues that are important to large slices of the electorate, all negatives for the GOP
Careful, Elliott: You're analysis is sophisticated but I think there may be a major flaw. In examining prediction-result patterns historically, your (small) set of elections (2010, 2014, 2018, 2022) balances three elections with a Democrat in the White House (2010, 2014, 2022) against one with a Republican (2018).
There's nothing inherently wrong with that, but it contributes strongly to the overall impression that the incumbency bias is nonpartisan -- i.e., bi-directional.
A strong case can be made, however, for regarding 2018 as highly atypical -- essentially a one-off, black swan event. Karl Rove, in a fit of pique over his displacement by Steve Bannon, took the 2018 election off. And I would submit that his talents -- i.e., his dark arts -- were sorely missed. (I'd also note, parenthetically, that the 2018 blue wave yielded a net gain of 40 House seats [and a *loss* of two Senate seats] while the 2010 red wave yielded a net gain of 63 House seats for the GOP -- an asymmetry not fully explained by organic factors.)
If 2018 had as much to do with who was navigating the GOP bus as with purported generic polling bias, then we're left with a series of elections in which pollsters missed to the left. And we can add to that the 2020 election, in which of 27 House contests rated tossup by Cook, the GOP won every single one (and also won 7 Democratic leaners while losing none of their own), and similar, if less egregious misses in 2016 and 2024.
So overall, the clear pattern in the computerized voting era is pollsters overestimating Democratic strength, making corrections, and overestimating Democratic strength yet again (pace 2018). Pollsters don't stay in business by erring repeatedly and consistently *in the same direction*! So it is understandable that Cook et al are lowballing Democratic prospects in these midterms. They have touched the hot stove -- over and over -- and are very wary of doing it yet again.
You can draw your own conclusions on why they keep missing in the same direction -- why, that is, there is a pervasive polling-to-votecount red shift, even after they tweak their sampling methodologies to account for it. Maybe there really is a "reluctant Bush/Trump responder" phenomenon that results in an oversampling of Democrats (but there is demographic stratification to handle that). Or, just spitballing here, maybe unreliable vote counts (from suppression schemes and/or vote-count manipulation) are a factor? Maybe all the red thumbs on the electoral scales have left the pollsters wondering what they keep doing wrong, when they're not doing much of anything wrong at all?
This is worth our consideration as we move toward an election in which Donald Trump has called upon his side (voters, administrators, operatives, vendors...) to "cheat like hell."
I mean, OK, but Karl Rove would have been fully in the game in 2006, and the Dem wave still happened that year.
Thanks for reminding me: https://codered2014.com/wp-content/uploads/2020/04/landslideDenied_v.9_071507.pdf.
What do we know about mail-in ballots relative to in person ballots? Have there been polls of already voted mail-ins one or two weeks before election day? (screening call is have you already mailed in your ballot and if yes you're in the survey. Or the question could be already voted/definitely will vote/ likely vote. Weighting the responses generates another complexity)
As is currently being noted on the talk shows voting is starting now and those votes won't change (in any significant way) from time of mailing to election day. Are there estimates of percentage of mail in ballots returned 4 weeks, 3 weeks, 2 weeks and 1 week prior to the election? Sentiment (or events) one week before the election won't change large percentage of the mail in vote.
Dammit, large GOP money well be hawking the “ewwww” factor, the fifth grade schoolyard stuff *really* hard starting in two weeks. It’s inconceivable that it will sway things — but it’s also inconceivable that it worked in 2024.
Maybe this time that will prove to be a well that had been gone to once too often? I’m still not completely over 1980, so we’ll see.
There are $6 diesel "billboards" that have been posted all over the country. There are $7/lb ground beef signs up in grocery stores. Elon can't by a $3 diesel billboard or reprice ground beef. Money can't buy you love and has significant limitations in buying votes.
I mean, that makes total sense. Selzer STILL hasn't lived down calling Iowa for Biden in 2024. Democrats are so het up at the moment that even if the magnitude is *slightly* off in the GOP's favor, pollsters are going to deal with a lot of shit. If they undersell it there's a lot less to lose.
The biggest X factor for November is whether half a billion of GOP SuperPAC spending this fall will shift close races towards the GOP.
That spending is up against the soaring prices of gas and diesel which will push voters towards Democrats.