US: Stop Believing State Polls

These Polls Are Crap

I really can’t believe we ended up in this place in 2022, but here is a sentence that I, almost uniquely amongst those who do election writing and modelling in America, am very qualified to write.

Stop Believing Dogshit State Polls.

Like, guys, I am the old LeanTossup guy who called Blue Texas and Blue Florida and wrote insanely wrong takes about Hispanic voters not actually swinging right. I was a blatant 413er. I was a Jaime Harrison can win South Carolina guy. I have more form on believing truly dogshit state polls because it said things that I agreed with than basically anyone who has the sheen of “independent” analyst (even if most Americans, not all that wrongly based on my US track record, think that I’m a delusional pro-Democratic hack). I genuinely believed TMac was gonna cruise to victory in November. Like, trust me – I don’t like it, but it’s pretty key to bring up all this history at this point, because y’all got to stop believing these shit state polls.

John Fetterman is not up 9% on Dr. Oz. John Fetterman is not up 9% on Dr. Oz on the same ticket as Josh Shapiro is only up 4% on Doug Mastriano. John Fetterman is not up 9% in Pennsylvania if Raphael Warnock is in a tie in Georgia, as ECU had yesterday. It is, plainly, just not true.

Ah, but what about the fact that Cheri Beasley was up 4% in North Carolina yesterday and Tim Ryan is in a close race in Ohio? Aren’t those three results fairly reasonable in comparison to each other? Sure, just like a thousand dogshit state polls made sense together in 2020. Remember when there was a West Virginia poll showing Trump up only 18%? Remember when there were close races in Missouri? Remember when CNN in the last two weeks had Biden up 5% in Ohio? If you don’t, that’s fine, but they are burned into my memory forever, because I did this in 2020 and holy fuck we need to stop.

I remember writing a column at some godforsaken early hour in the morning on maybe 2 hours sleep about an ABC/WaPo poll of Wisconsin, remember this? Final week, I think, and it had Biden up 17%.

He won it by less than 1%.

Stop Believing Dogshit State Polls.

Ah, but the state polls in Georgia and Virginia were good, so we can trust these, right? No, because samples of two are dumb and the fact that polls were good in 2018 lulled us all into a false sense of security about 2020, but also because those California recall polls were dogshit, and so were the New Jersey polls. And nothing should make us confident that the state pollsters have adjusted after 2020, because it looks like there’s a huge disconnect between the state polls and the national polls.

Basically, after Virginia the Generic Ballot lurched right, in part because pollsters with bad Biden approvals starting polling the Generic Ballot as well, but partially because the polls started to pick up huge Republican enthusiasm advantages. Whether that was an intentional choice to boost it because of Virginia, or those pollsters made that choice due to their samples, I have no idea. But the Approval and Generic Ballot polls nationally all have Republicans as more enthusiastic to vote – a trend that is matched by the historical pattern of out parties in midterms, if you care about such things.

All these state polls have big enthusiasm advantages too, but for Democrats. Now, could it be a Roe effect? Maybe, but it’s not – it’s Democrats like answering polls more than Republicans, and they’re not getting representative samples of Republicans. It’s very simple to see that the same polling error that had Biden winning Pennsylvania by high single digits for months and had Ohio close, depending on the pollster and whether it’s a red or blue sponsor is why Fetterman is up 9% and Ryan is close. Democrats getting burned by North Carolina polls not matching reality? Where have I heard this, except every Presidential and Senatorial election since Obama and Hagan actually won it in 2008.

If you take every poll with May or June field dates, take out the Mark Kelly up 15% in Arizona Senate, the Catherine Cortez Masto is up 23% in Nevada, the weird ass poll with Val Demings ahead in Florida, and the Blueprint Texas poll with Greg Abbott up 19% - so, four outliers that are clearly just not correct at all, and 3 of which are outliers that would skew this average towards Democrats – the average swing right from Biden’s 2020 result in the state is 1.7%. If you take out the most R leaning and the most left leaning outliers – the “Olympic Average”, the way they judge Olympic diving and figure skating and what not – the average swing right is 1.8%. The median swing is a gaudy 2.7%, but all three of these state polls based calculations point to a national environment where Democrats win the popular vote by 1-3%, and are somewhere between 35/65 and 60/40 to win the House.

If you use my national polls model to calculate a Generic Ballot – based on correcting approval polls for their flaw that many Black and Hispanic voters are Biden Disapprove, Dem 2022 voters – and I have the swing at 9.2%, a National environment of R+4.7. I have never seen two metrics both in theory equally valid pointing this far in opposite directions ever – either my own, or anyone else’s in public. And the thing is, the metrics are both right, in a sense. They both accurately convey the current state of the thing they’re supposed to analyse.

If you take the state polls in aggregate right now, Democrats are going to hold almost all of their Governorships, they’re looking at Senate gains, and the swing in the Generic Ballot to Republicans is probably smaller than what Republicans need, especially if the 2012 to 2020 trends in white suburban districts continue. It’s certainly smaller than what Republicans would need to feel comfortable of winning the House, as they do.

Like, the metric is doing its job of giving me a quantifiable assessment of what I was subjectively thinking for months now, that these state polls are just hilariously out of whack from the national polls. But the metric is also wrong, because these state polls are dogshit. I don’t have a better word for them.

There is no way Cheri Beasley is up 4% right now. There’s no way John Fetterman is up 9% right now. There’s no way Tim Ryan loses by less than 5%. Listen to the former delusion fuckwit who fell down this rabbit hole last time. There is nothing but doom and pain for you. Stop. Please. I’ve been here before, I’ve seen this room and I’ve walked this floor. It all ends in tears and having to profusely apologizing for believing dogshit polls.

Look, you want to sell me that Fetterman might be a small favourite now? Maybe I can get there, but I can’t on Beasley, no matter how many polls you release showing her up. If you want my Senate take, I oscillate on whether it’s 49/51 or 50/50, because I flip on any given day on whether Blake Masters being a horrible candidate is enough to overcome the fact that Biden won it by a shade over 10k votes.

I think there are four Senate seats that are meaningfully in play – Nevada, Arizona, Georgia, and Pennsylvania – and Democrats need to win 3 of them. I think they’re fine in Nevada, so that’s seat 48. I think Warnock gets there in Georgia, especially given Herschel Walker is suffering from undiagnosed CTE and literally does not know where he is or what he’s talking about at time (and has a secret kid while his son is ranting on TikTok about how America has a crisis of absentee fathers), so that’s 49. I think between Arizona and Pennsylvania Democrats probably can squeak over the line in one of them, but I have the GOP small favourites in both as of this exact second.

But all of that actual nuance is not the point of this column, and I don’t really care if you think I’m some Dem shill for thinking that a Hispanic woman will do better with Hispanics than a white dude who spent 0 time trying to appeal to the community and has staffers who all think Latinx is reflecting the community better as opposed to really telling on how white you are.

The point of this column is very, very simple: a 413er is telling Democrats to not believe these state polls. They are actually, genuinely, legitimately insane. Ignore them. They are wrong, they are delusionally wrong. Stop. For the love of all things holy, stop. It’s not true, Democrats aren’t doing this well, and everyone needs to calm the fuck down and stop believing crap just because you really, really want it to be true.