2022 Generic Ballot / Recruitment / Fundraising / Ratings Megathread

[ElectionsGuy]

Maybe some of these candidates should wait to see what their districts are going to look like first before running?

[ElectionsGuy]

Christy IN!

https://twitter.com/ChristyforCA25/status/1333806001921826816

Ah geez! She just lost a very winnable House race. We don't need a Martha McSally of CA-25.

She underperformed Biden by around 10 points. She's not Martha McSally, she's more like Martha Coakley.

[ElectionsGuy]

Best place to post this...Congress is somehow not unpopular? Clearly there has been a opinion shift among the Democrats.

https://twitter.com/USA_Polling/status/1390755159542751241?s=20

That poll is a huge outlier. They also have Biden approval at +21%, which is way too high. Although Congress does now have the highest approval they've had since 2009. Democrats clearly now like Congress that they've got both chambers. Republicans though have always disliked it even when they had control, as much of late Obama/early Trump was spent below 20%.

[ElectionsGuy]

The oversampling Rs is fine, giving Dems a scare. The tie is also in battlegrounds even with this oversample. Hate these multiple state/combined district polls, such a weird sample.

A battleground phone poll of 1,000 registered voters, with an oversample of 500 Republicans was
conducted by phone April 27-May 3, 2021 from a voter-file sample. 67% of respondents were reached
on cell phones in order to accurately reflect the American electorate. The margin of error is +/- 3.5
percentage points.

4 South/Sunbelt States = AZ (+5), FL (+2), GA (+5), NV (+1)
7 Blue Wall States = MI (+1), ME (+3), MN (+1), NH (-), OH (+3), PA (-), WI (-)
13 Additional Frontline Congressional Districts:
• CA-39(+3), CA-45 (+3), CA-48 (+1), CA-49 (+4)
• NJ-03 (+3), NJ-05 (+1), NJ-07(+1), NJ-11 (-)
• TX-22 (+4), TX-24 (+2), TX-32 (+1)
• VA-02 (+1), VA-07 (+3)

Oversampling of Republicans (or any other group) doesn't mean [at all] that the poll's skewed. They don't give more weight to R. They just wanted to decrease MOE *inside* the group to study Republicans' attitudes towards Trump etc.

Oversampling is going to be critical to any sort of rebound in polling accuracy. And it shouldn't necessarily be Republicans, but the demographics associated with non-response, like region, education, etc. Also need a variety of sampling methods since live caller is nearly obsolete at this point.

[ElectionsGuy]

Democrats surge into a ten point lead!
https://twitter.com/electsworld/status/1395746724879216646?s=21

A lot of the GOP's optimism comes from historical trends, not actual polling for 2022.

To be fair, polling is unreliable at best. Answering polls have become associated with being a Democrat.

Polls struggle when Trump is on the ticket. When he's not, they're largely pretty accurate.

That's why in the last midterm of a Democratic president, polling accurately reflected Hagan getting re-elected, Cotton winning by single digits, Ernst narrowly beating Braley, Walker narrowly beating Burke, Quinn getting re-elected, Brown beating Hogan, Georgia races going to a runoff because they were so close, and last but not least the national house vote going narrowly Republican after being a Democrat lead or roughly tied for most of the year and many "reputable" pollsters showing Democratic leads.

I'm sure we'll get a lot of THIS ISN'T GOING TO BE 2014 responses, just like the THIS ISN'T GOING TO BE 2016 responses in 2020 I became very familiar with, and like the THIS ISN'T GOING TO BE 2010 responses back in 2014. And, hot take here, we'll get THIS ISN'T GOING TO BE 2020 responses in '24.

[ElectionsGuy]

It is very disappointing that people who should know better (harry enten, nate silver, cohn, etc) are still retweeting QPAC as if its polling is adequate.

they have burned their reputation. They don't deserve our trust until they have an accurate cycle again

They don't care. These people are Democrats so they like the numbers they put out. It's that simple. And at this point, it's hard not to come to the conclusion that some of these pollsters (like The Hill, Reuters) are not purposely inflating their numbers for the purpose of Republican demoralization or to feed their audience what they want. 

If polling was this systematically slanted against the Democratic Party, these poll watch prognosticators would not be behaving the same towards them. They would be getting even worse treatment than Trafalgar did last year.

[ElectionsGuy]

The problem isn't Biden, it's the Dems in Congress who consider themselves "the future" who have an ideology far to the left of the median voter. It turns away a portion of very gettable people for the Dems who aren't interested in AOC's version of America.

The thing with this argument is that Republicans do the same thing, and even worse, with members like Greene and Gaetz, but for some reason, pundits and others don't see them as arbitroses like AOC?

Do you not watch MSNBC, or CNN? They are constantly lashing out at Greene and Gaetz and pretending they represent the entire Republican Party.

[ElectionsGuy]

Guys the Dems are nervous and think they're losing ground but they also want you to know their agenda is 'ultra popular'. Make sense?

[ElectionsGuy]

"Forty-seven percent of registered voters prefer a Democrat-controlled Congress, while 46 percent want Republicans in charge. (That’s down from the Democrats’ 5-point advantage on this question in April.)"

https://www.nbcnews.com/news/amp/ncna1277368?__twitter_impression=true

Well, democracy was fun while it lasted.

Democracy means your people can be voted out too.

[ElectionsGuy]

Imagine believing the Democrats are going to do 5 points better than in 2020 in a midterm when the incumbent Democratic president is nearly 10 points underwater. lo freaking l.

[ElectionsGuy]

Imagine believing the Democrats are going to do 5 points better than in 2020 in a midterm when the incumbent Democratic president is nearly 10 points underwater. lo freaking l.

This is the generic ballot thread, I'm just posting polls as I see them, they're not an implicit endorsement of them on my part (IMO this one has way too many undecideds).

That said, IIRC you said the exact same thing about a Stansbury +13 poll, in the exact same rude, haughty tone, so maybe your feelings on polls don't always match with reality.

I don't think that's true (I think the poll was better for Stansbury than that) but I remember saying you guys are setting yourselves up for egg on your face if you believe a race is certain based off one poll. And I stand by that, despite it not coming into use for that election. This is about the generic ballot, a much more broad and certain measure than a congressional special election.

By the way the "believe" part wasn't directed at anyone in particular. It wasn't meant to be rude, it was a light-hearted post actually. But we know people out there will just believe any polls that are out there no matter how unrealistic they are. That's why I said "imagine".

[ElectionsGuy]

https://twitter.com/mattholt33/status/1448250267934330883

I'd never noticed it before, but does Marco Rubio have really big ears?

Yes. They're massive.

[ElectionsGuy]

Just so you know, 538 has Democrats averaging the exact same 3 point House lead they got last year. You could easily adjust it to be a narrow Republican lead, or just count on the generic ballot being a lagging indicator compared to presidential approval. So far, it's better than the 39% and 7 point deficit 2018 Republicans were recording at this point, but that doesn't mean they'll retain the House.

The generic ballot polling early on has a tendency to lag behind the shifting environment. I remember throughout much of 2013 and early 2014 Democrats led the generic ballot quite often, and only in the last few months did Republicans take a narrow lead, and then they went on to win by 5-6 points. In 2018, one of the only years recently that hasn't underestimated Republicans, Democrats had a softer ~6 point lead before getting into 8-9 territory before the election. In 2016, Democrats led by 3-5 throughout, and then it narrowed toward the end to a 1 point lead and Republicans won it by 1.

Given the polling for 2020 had Democrats cruising to a slightly smaller than 2018 margin, a Democrats +3 aggregate is not good. They're pretty much on track to lose the nationwide congressional vote. But if it doesn't get better for Republicans than this in next year of polling, I think we're looking at another historic miss. Especially as some pollsters still have a laughable +8 at this point, that doesn't make me confident that polling is getting better, in fact it may be continuing to get worse.

There's no election, none, in the last 30 years or so where the incumbent's party did better or as well as the previous presidential election when the incumbent's approval rating is underwater. And when you pin down who's undecided, it's very disproportionately independents who disapprove of Biden. Gee, how do you think they're going to break eventually? The people who still approve of Biden are on the D train no matter what, but the people who don't only have one political party to blame and one other to choose from.

[ElectionsGuy]

With what just happened in Virginia and New Jersey, I don't buy any poll that currently has Dems in the lead for the generic ballot.

They're behind, it's just a matter of how behind they are, and whether they can improve this or not before 2022.

[ElectionsGuy]

R+6 type of environment

I am thinking that 2022 will result in the Republicans possibly having a clean sweep of around 80% of the Senate races and maybe having a net pickup of around 100 House seats.

Maybe if Republicans had previously been wiped out like during the great depression. That's completely impossible today, with 213 current seats + 160ish deep blue districts that never go R in any scenario.

[ElectionsGuy]

The state of polling in 2021

[ElectionsGuy]

Nunes can still win in that seat, but it'd be harder to hold after '22.

[ElectionsGuy]

The generic ballot polling at this point is more random than 2014's polling was. R+10 is still closer to reality than Economist's D+7, or CNN's D+5. Given we saw roughly R+7/8 results in VA/NJ, and Biden's approval in RCP was -8 at the time, we now have a decent idea of what's realistic and what isn't.

[ElectionsGuy]

https://twitter.com/PollsandOdds/status/1461013026023383048?s=20

In no scenario do Independents go Republican by 10 and Dems win overall by 4. YouGov's polling has and continues to misrepresent the electorate.

[ElectionsGuy]

It seems Marist, unlike Quinnipiac, wants to continue embarrassing themselves.

[ElectionsGuy]

It seems Marist, unlike Quinnipiac, wants to continue embarrassing themselves.

While Marist has always been D-leaning and I don’t really ‘trust’ this poll, it’s crazy to throw out a poll from a reputable pollster. Maybe Ds are just seeing some regression back towards the mean (as literally always happens) and simultaneously this sample is a bit D-friendly.

The tendency of this forum to dismiss polls that’s don’t align with their preconceptions rather than try to understand them is pretty silly.

Ah yes, the "reputable" Marist poll. Let's go back, shall we.

NBC/Marist

US: Biden +11 (D+7)
AZ: Tie (0)
FL: Biden +4 (D+7)
MI: Biden +8 (D+5)
NC: Biden +6 (D+7)
PA: Biden +5 (D+4)
WI: Biden +10 (D+9)

Average Error: 5.3
Average Partisan Error: D+5.3
538 Grade: A+

Nate Silver's favorite pollster for a long time.

That doesn't include their bullseyes in 2018, which included McCaskill +3 for MO Senate, Evers +10 for WI governor, Sinema +6 for AZ Senate, Blackburn +5 for TN Senate (Bredesen +2 in August, btw), among others, in a year where most people here believed the polls were accurate, proving that 2016 was a fluke for polling, and that 2020 polls would be mostly accurate.

More importantly, we just had elections a few weeks ago, where Biden's approval was -8 at the time of the election in both 538 and RCP's averages, and Democrats lost Virginia and narrowly won New Jersey, with incumbent and former incumbent Dem governors running and conservative "Trumplican" Republican candidates. Averaging the swings from 2020 in both states would suggest around R+8 for the country. D+5 is 13 points away from that, way outside any reasonable polling error, or any reasonable explanation for why Republican governor candidates would do so much better. Let's remember 2020 for the House was D+3. This poll is claiming Dems have a roughly equivalent or slightly larger lead today, right now, with Biden's approval being roughly the same as November 2nd, 2021. And, it would be one thing for a poll to claim a narrow Dem lead with lots of undecideds, say 42% to 40%. But for this poll to claim 46% would vote Dem already, with 13% undecided, says a lot of things in my mind about the poll, but none of which are that the poll is accurate.

If we had polls claiming R's leading in 2018, everyone here would be laughing them off. And they should've been laughed off, because we knew that was nowhere near reality based on multiple indicators. People should do the same here even if they want to believe it.

[ElectionsGuy]

https://twitter.com/HotlineJosh/status/1468289128529899527?s=20

As much as the generic ballot has gotten better for Republicans in the last few months, these numbers are still underestimating R's. Like it's pretty telling that 57% disapprove in this poll yet only 44% are going Republican. There is a point in the electorate where no matter how low Biden's approval goes a certain percentage will just vote Democratic anyway, however, that kind of separation (44-57, vs the 41-41) is telling me that those are people who don't necessarily like R's but they will vote for them and they're hesitant to say so. We didn't see this kind of thing in 2017, it was blue wave territory immediately.

[ElectionsGuy]

How does YouGov/the Economist actually have a D +9 generic ballot? That is such a glaring error that one would think they would make efforts to change their methodology to achieve more realistic results. Their approval numbers for Biden are reasonable, so why is the GCB so extreme. How are they sampling so many Biden disapprove/congressional Dem supporting people?

Is it worse than the Rasmussen R+10 from a while back? I don't think so. Just because it does not line up with your priors it does not make it wrong. That having been said D+9 is probably an outlier IMO but like all polls it should be thrown in the average.

Yes, it is. Nothing right now supports the idea of Dems even slightly winning the national popular vote.

[ElectionsGuy]

How does YouGov/the Economist actually have a D +9 generic ballot? That is such a glaring error that one would think they would make efforts to change their methodology to achieve more realistic results. Their approval numbers for Biden are reasonable, so why is the GCB so extreme. How are they sampling so many Biden disapprove/congressional Dem supporting people?

Is it worse than the Rasmussen R+10 from a while back? I don't think so. Just because it does not line up with your priors it does not make it wrong. That having been said D+9 is probably an outlier IMO but like all polls it should be thrown in the average.

Yes, it is. Nothing right now supports the idea of Dems even slightly winning the national popular vote.

I agree that D+9 doesn’t make sense, but Biden’s approvals have rebounded marginally in the last month.

If this is a rebound, I don't want to know what a falling out is.

That week or so where it went down to -6 was also around the time the Biden administration pressured the media to be even more favorable to him, and a bunch of the media polls that are usually better for him came out.

[ElectionsGuy]

Has anyone else noticed that the Democrats have been doing pretty well in generic ballot polling recently? On January 5, 2018, the Dems were up 11 points on the RCP aggregate. The GOP is up by 1.6 right now. 538 had the Democrats up 10, but the Republicans are up by only 0.5 at the moment. The midterms are obviously still going to swing R, but it's probably not going to be some 2010 or 2014 redux like some doomers on here are heralding.

That could be a good point but there is a possibility that republicans will overperform their current polling or gain in their polling lead as the year goes on

Just a reminder that the generic ballot was D+1 when Glenn Youngkin (and every other statewide Republican) won narrowly in Virginia and Phil Murphy got 51% in New Jersey.

Also, Dems were "polling" at D+7 on the generic ballot in 2020, when they actually won by 3. This is going to be a historic polling miss if this average stays.