One thing I've been experimenting with since our GOP oversample this summer: weighting our polls by each partisanship subgroup, which has the consequence of ensuring that each subgroup used in weighting has the right number of Dems and Republicans The main downside is that our estimates for self-reported education by voter file party are modeled, and that's something I've had pause about. I'm gradually getting more comfortable with it, as the party x edu tallies for the typically weighted sample seem consistent with subgroup

I happened to be looking a lot at Pew data last few weeks, even before this most recent partisanship study, so I wanted to share a few interesting observations about trends I noticed in their data
pewresearch.org/politics/2024/… One thing I noticed: subtle but persistent, multi-year differences between the partisan splits by demographic on the Pew ATP -- the mostly mail-to-web panel they use for this study -- and the Pew NPORS study (the one-off high-incentie mail survey they use for weighting the ATP

*Tosses meat into cage* nytimes.com/2024/02/14/ups… A few outtakes:
-- By our (rough and preliminary) estimates, this looks to be yet another zero-persuasion (off Biden '20) special. We'll have to see final vote history, but at least in Nassau it looks just as we'd expect given the party reg turnout

lol well did I get replies to this!
A few notes on special elections / clarifications

https://x.com/Nate_Cohn/status/1747441818612945200?s=20

1) Specials are driven by turnout. The data is unequivocal as long as I've been looking at them with our rich data, going back to 2017. That should not be remotely surprising, as the as the people who know about/vote in specials are highly parstisan -- just like all of you!

We have a new NYT/Siena national survey, and it's an interesting one -- with the public sympathetic toward Israel but disapproving of Biden on the issue and split on whether Israel should continue military operations.
It also has interesting '24 numbers...
nytimes.com/2023/12/19/ups… Trump leads 46-44 among RVs, but *Biden* actually leads in our first measure of the likely electorate nationwide, 47-45.
The split is driven by a huge gap in vote choice by turnout history: Biden+6 among '20 voters; Trump+22 among 2020 nonvoters
nytimes.com/2023/12/19/ups…

One thing I've seen over the last few days: a lot of people asserting that, in a variety of different ways, pre-election polls aren't very useful for demographic subgroups
I have to completely disagree. Stepping back, it's my long-standing view that the pre-election polls are the best basis for post-election estimates -- and, in particular, better than exits.
For ex, all these estimates were based on pre-election polling:
nytimes.com/2016/06/10/ups…

One year out, Trump leads Biden.
The poll finds Biden is pretty badly damaged, but that there's still a path for Democrats to win.
Whether Biden can capitalize is the big question
nytimes.com/2023/11/05/ups…
Since the last election, Biden lost the advantages that let him beat Trump. Simply put, he's no longer the broadly appealing, well-liked candidate with an advantage on favorability and so on who beat Trump four years ago

Trump 54, DeSantis 17 in the first NYT/Siena poll of the cycle. No one else is over 3%
nytimes.com/2023/07/31/us/… The poll amounts to a fairly comprehensive takedown of the DeSantis campaign theory of the race, with Trump enjoying strong advantages on DeSantis' main arguments (electability, getting things done), while his issues (woke) seem to lack much punch with conservative voters

nytimes.com/2023/06/01/ups… Quick outtake for the scatterplot lovers. Black is 2012; red 2016; blue is 2020. Data uniformly adjusted to match election result

We'll undoubtedly start to get some post-indictment GOP polling over the next few days. My preemptive plea: please be sure to compare a polls to prior results by the same pollster This cycle, pollsters have split in extremely unusual ways about both the state of the Republican race and the broader trendline.
There's a real risk of being seriously misled without understanding how pollsters have been behaving.

I've seen two articles on recent electoral trends this week. Both worth reading, but I'd like to add a sidebar:
Our ability to measure electoral trends has degraded significantly over the last few years.
washingtonpost.com/opinions/2023/…
nytimes.com/2023/03/22/opi… There are two main reasons:
-- Changing methodology has undermined comparability of data across time, even from the same source (exit polling in '20 is way different than '16)
-- 2020 polling misfire creates serious questions about validity, even if it was comparable

"Siena College/The New York Times Upshot already had an A+ grade, so it didn’t get a ratings boost. Still, its stellar performance did push it past Selzer & Co. for the distinction of most accurate pollster in America"
fivethirtyeight.com/features/2022-… By my count, the Times/Siena polling this cycle was the most accurate for any pollster with >10 polls in the FiveThirtyEight dataset going back to 1998. Pretty amazing given all of the challenges in polling today

Do you think turnout A or turnout B is better for Democrats or Republicans? (poll below) Do you think turnout A or turnout B is better for Democrats or Republicans?
(redoing question bc i'm bad at this)

Happy Election Day
A few morning thoughts on the Georgia Senate race, which is close and yet seemingly fairly clear
nytimes.com/2022/12/06/ups… I'll be honest, I think it's pretty hard to tell a story about why Walker wins this election. That's not to say it's impossible -- anything can happen. I think it would just be hard to explain how and why, given the candidates and the hard results last month

Well ... we've been on the "cusp" of a House takeover for a while now.
While we wait, here's a quick rundown on how the post-election count has gone in the outstanding CDs that have a chance to be called sooner (notice that this is a relative term) than later CA27:
--On 11/9, GOP led by 15!
--Now it's down to 8 points, as the Democrats have won the post-election count by 9.
--But to win, Dems would need to win what's out by 19.
Yesterday was an even split.

With the count almost done, the NYT/Siena poll appears to have had its best year yet: an average error of 1.7 pts with 0 bias
(the national popular vote is estimated with imputation in uncontested races, who knows where that lands) 🤷‍♂️
IDK.

https://twitter.com/Wertwhile/status/1592246008674545666?s=20&t=OFjU7RJ6uom5M8ydfm7PoQ

No projections in AZGOV, but make no mistake: Lake didn't get the tallies she wanted and probably needed out of Maricopa.
It will presumably close further, but Lake doesn't have many batches left. Each time she falls short, her target in the outstanding vote gets higher. A projection means something very close to absolute certainty -- say, > 99.9% chance.
There's a lot of room for something to be very likely (90%, 95%, even 99%!) and still not quite merit a race call.

https://twitter.com/keithedwards/status/1591970960726700032?s=20&t=9_OgQ7AoP2RaPadNn6JO9A

A look at how the late mail votes are breaking in the uncalled House races.
So far, it's decent for Dems in California, but not enough to be favored to go ahead without additional strength
nytimes.com/2022/11/12/ups… The thing that keeps Democratic hopes alive is that they *are* gaining in the post-election vote in most districts. At the moment, just not by enough

GOP leads in 221 seats; Democrats need to flip at least four where the GOP leads
The three likeliest targets, to me, seem like CA21, AZ06, CA13.
After that... it's murkier. Lots of vote left in CA27 and CA45, but Dems not yet doing what they need
nytimes.com/interactive/20… sorry, meant CA22 not CA21. on the wrong decade's map still

https://twitter.com/JesseLehrich/status/1591257807570890753?s=20&t=D8qP9XdVvlANG_xRsBA3TQ

Maybe I'm naive, but I'm a little surprised by the amount of griping about the polling that I'm seeing.
The polls did pretty well! The 'traditional' polls did *really* well. Doesn't get much better I do think there were some interesting challenges in this election, which maybe I'll write about:
One challenge: the state-level variation in the results (FL/NY bad for Ds, v. PA/MI great, etc.) made it very hard to synthesize seemingly conflicting data.

I would think so. The Dems holding up well in the post-election mail increases their chances a lot in a place like WA-3, where on Election Night the needle would understandably wonder whether it's 'blue mirage' until shown otherwise

https://twitter.com/BroChillerman/status/1590733303552049152?s=20&t=-xyzoHS9FfhaptmrktpfxQ

Many asked why the needle turns off. Two related reasons.
One: once our turnout estimates (200k votes left, +/- 50k) get replaced by hard facts (173k votes left), we're put in a spot where we're always inaccurate or forced to be constantly updating based on reported data.