Virginia: Leave Me (Like You Found Me)

An Ode For Sanity

When was the last time US polls were more instructive than just assuming not much would change?

I’ve been thinking about this question for about 6 hours, and I think the answer is the 2014 Iowa Senate race. In 2018, the polls mostly missed by exaggerating the distance between the 2016 result and the eventual result, the 2020 polls were absolutely putrid, and in 2016 both a nihilistic “nothing will change” approach and the polls were proven wrong, so then it has to be 2014. Even broadly that year, just defaulting to very basic knowledge - “red states will elect Republicans” - got you 7 of the 9 Senate seats correct, but Iowa was the one where the polls actually added some value. Except, it wasn’t the polls, it was Selzer, and Selzer alone.

Funny how that works.

…

Had your operating assumption about the 2020 election been “at best we’ll see changes at the margin”, you’d have done a lot better than anyone else did, right? You’d have had the three Upper Rust Belt states flipping, but not safely, you’d have had Susan Collins in a more competitive race than the polls did, you’d have been more sceptical of Cal Cunningham the entire time, and you’d have outright dismissed the lunacy stories people told about South Carolina and Montana and Alaska and Kansas (of which I am exceptionally guilty). Now, maybe you’d still have had Florida flipping, and Georgia probably would have seemed a bridge too far, but in terms of the broad picture, you’d have been a lot more right than basically everyone.

If you had taken this approach to the 2018 Governorships, you’d have been correctly pessimistic about the chances Billie Sutton could win in South Dakota or Richard Cordray could win in Ohio. You’d have not bought Democrats winning the Iowa Governorship, and you’d have seen the Abrams narrow loss as well within the realm of possibility. Again, on the margins things got wonky - Gretchen Whitmer winning by the amount she did would have been wild to contemplate, but still, a lot of the landmines of 2018 would be missed.

At some point, we need to have this conversation because these Virginia polls are just flat out bad. They’re horrendous, and they’re horrendous even if McAuliffe somehow loses. I’m sorry, they just are. Is Glenn Youngkin somehow outrunning Trump by 24% with non-degree whites (as Monmouth claims)? Is Youngkin suddenly going to do nearly 40% better on margin with Black voters than Trump (as both Emerson and Cygnal claimed)? Do either of those things actually seem believable to you? More importantly, would they be believable to you if they were inverted, and we were talking about Jaime Harrison? Oh, I know they’re not, because I still get the tweets about that shit.

This isn’t about Virginia, but it is. On Saturday I wrote that the GOP were making very small gains with white educated voters and that they’re not making anything resembling the gains they’d need to make this race, how would you say, interesting or competitive, and yet here we are. This isn’t about Virginia because this is about good faith and not, intellectual honesty and not, and the lessons from a very long 2 and a half years of being someone with something resembling a public platform.

If I ever write that Tim Ryan’s a favourite in Ohio, go ahead and never listen to another word I have to say, straight up.  If I ever actually write or say the words “Charlie Crist is more likely than not to win”, just unsubscribe to this site and my podcast and unfollow me on Twitter. This is about intellectual honesty and good faith, and if my argument in Virginia is partisanship and polarization are the overriding factors, and dodgy polls should be dismissed, then I have to hold that line when the dodgy polls favour the party I want to win. Otherwise, it’s just bad faith.

I really like being right - like, I really like being right, probably to an unhealthy degree. The night the UK elected Boris with that huge majority, I celebrated wildly, because I was right. The fact I hated Boris wasn’t relevant - I was right. Same thing with the Australian election months before - I had bet a decent chunk of change on a series of bets wherein the Australian right was going to have a good night, and they had an even better night than I ever could have asked for. It was an immaculate night (morning, technically, I think?), and I was overjoyed - because I was right. I want to be right more than I want to predict good things for the political persuasions I support. And I am still convinced McAuliffe is going to win, and win easily.

I’m publishing a novel in December (apologies in advance, but it’s going to come up a few times here), and it’s funny to think back to the process of starting that story having finished it, and more than that, having written a pair of other novels in the months after. It’s a good story, at least to me, but what’s funny about it to me is how a completely different story could have been written with the same characters, but with slightly different choices. Seemingly small decisions about backstories inform so much of the story, and in the same way, minute distinctions about how to read data can completely change the way a race is viewed. In the same way that the story I wrote is, in my view, the best story that could be told, my view of the data is in my view the best way to look at it.

The title of this piece comes from the Wilco song of the same name, and it’s playing as I write this piece. If your assumption about this race is that nothing has changed, you’re almost assuredly correct. McAuliffe won’t win by Biden’s 10% margin, probably, but he’ll win comfortably, and easily, and the right way to think about this race is to do nothing, and leave Virginia like you found it.