Share this page!

Enter URL or ID to Unroll

You can paste full URL like: https://x.com/threadreaderapp/status/1644127596119195649
or just the ID like: 1644127596119195649

How to get URL link on X (Twitter) App

  1. On the Twitter thread, click on or icon on the bottom
  2. Click again on or Share Via icon
  3. Click on Copy Link to Tweet
  4. Paste it above and click "Unroll Thread"!
  5. More info at Twitter Help
Nate Silver Profile picture
Nate Silver
@NateSilver538
New Book, On The Edge, #5 NYT Bestseller! https://t.co/WeCLEOd4Be
621 subscribers
Jul 3 • 5 tweets • 2 min read
We ran one more update, just because we don't want to have to worry about it on July 4. Trump national polling lead up slightly further to an even 3 points after a couple more national polls.
natesilver.net/p/nate-silver-…
Image The model DOES update its STATE polling averages based on trends in national polls (and polls in other states) but it's probably being a little conservative there (i.e. it wants to see more state data); a little bit of an imbalanced ratio of national to state polls this week.
Feb 10 • 5 tweets • 2 min read
It's definitely one of the more fortuitous accidents of timing I've had as a writer. On Tuesday, I wrote a piece saying Democrats use "But Her Emails" to deflect legitimate criticism. And that's exactly what they've done since the special council report on Thursday...
Image
Image Biden's age is not at all comparable to Hillary's emails. It is a much more important issue. He wants to be president until he is 86 years old! Voters ratioanally think it's important. I criticized #ButHerEmails early and often. This story is not the same.
natesilver.net/p/not-everythi…
Jan 12 • 4 tweets • 2 min read
I'M SORRY BUT YOU DID A MISINFORMATION SANDER! YOU'RE ONE OF THE BADDIES! You've routinely spread misinformation about the scientific consensus on COVID origins. The fact that you can't acknowledge this why the concept is incoherent. Half the reason the Team Misinformation people bug me is because it's just so obvious what they're doing, taking genuinely contentious discussions and stigmatizing the positions that don't match their politics with the thinnest imaginable reeds of expert authority.
Feb 2, 2023 • 10 tweets • 3 min read
🧵1/ Our biennial forecast self-review is out! There’s lots of detail in the story, please check it out. We think it’s really important to do this. It’s also one of those years where it may clear up some misconceptions.
fivethirtyeight.com/features/how-o… 2/ Polls (in the aggregate) and forecasts had a good year. Polling averages did ~not~ predict a red wave. They showed a highly competitive race for the Senate and below-average (by historical standards) GOP gains in the House, though with much uncertainty. Image
Nov 9, 2022 • 4 tweets • 1 min read
This is cool. GOP currently leads 220-215 based on called races + races where they're currently ahead. But, quite a few are uncertain; some key ones below. 4 races where Ds currently lead but the Upshot Needle (as of 4AM last night) had Rs favored:

AZ-1
CA-41
CO-3
WA-3

1 race where Rs currently lead but the Upshot Needle (as of 4AM last night) had Ds favored:

MD-6
Nov 8, 2022 • 5 tweets • 1 min read
Republicans have a 59% chance of winning the Senate, according to our final Deluxe forecast. It's closer in our alternative models: R chances are 51% in the Lite (polls-only) forecast, and also 51% in Classic (polls + fundamentals but no expert ratings). projects.fivethirtyeight.com/2022-election-… In the House, Republicans are considerably more definitive favorites: 84% in Deluxe, 82% in Classic and 75% in Lite. Still, you shouldn't round their chances up to 100%. It wouldn't require **that** large of a polling/forecaster error for the House to be competitive.
Oct 28, 2022 • 4 tweets • 1 min read
Some qualifications on this, and more in some races than others, but our Deluxe model expects Republicans to outperform their current polling by ~2 points or so in the average Congressional race.

fivethirtyeight.com/features/where… Here is the comparison in the Senate, for instance.
May 19, 2022 • 6 tweets • 2 min read
Seeing a lot of this sort of sentiment. "Well, there's been no big poll shift after the Roe leak, which means no big effect on public opinion". But I think it's misguided for a couple of reasons. 🧵 1. Polling is noisy, and the generic ballot is particularly noisy. If e.g. Roe shifts the political environments toward Ds by a net of 2 points, that's a fairly big deal; could save them a couple of key Senate seats. But that could take a while to show up in our averages.
Apr 28, 2022 • 5 tweets • 2 min read
Folks, it's not that hard, the left has moved to the left in the US *and* the right has moved to the right. You can, of course, complicate the story as much as you want, including by noting that the *Democratic Party* is not synonymous with the left and the Republican Party is not synonymous with the right and the *GOP* has probably radicalized more than the Democratic Party.
Apr 9, 2022 • 4 tweets • 2 min read
Very very bad when the range of final polls is this tight. Almost certainly a sign of herding (pollsters trying to avoid standing out rather than reporting what their data actually shows). Tweet I was responding to got revised (@Taniel is an excellent follow BTW) but this is still a very tight range of results that likely reflects herding.
Apr 9, 2022 • 11 tweets • 3 min read
Inspired by the recent @Herring_NBA @ZachLowe_NBA All-Defense podcast, a 🧵 on why advanced NBA metrics (and in particular 538's RAPTOR) like Nikola Jokic's *defense* quite a bit. First, rebounding. Defensive rebounding is an important part of defense. Jokic is second in the league in defensive rebounding %.
Jan 21, 2022 • 4 tweets • 1 min read
It seems to me like the "right" way to frame the critique of Biden's legislative agenda is not that it's too ambitious or too left-wing per se but that voters want him to address *acute* problems whereas something like BBB tends to address *chronic* ones. "Theats to democracy" is a more complicated case in that there are both chronic and (extremely!) acute problems. But the legislation proposed to date has also tended to be geared toward addressing the chronic ones.
Jan 18, 2022 • 4 tweets • 2 min read
This is basically wrong, as a blanket statement. Some ideas championed by progressives are fairly popular in nonpartisan polling (voting rights is one example). Some ideas (such as on educational policy) are fairly unpopular. Some are a mixed bag. Biden's signature policy priority, Build Back Better, is in the "mixed bag" category. Its popularity is lukewarm, both for the package as a whole and for the child tax credit in particular.

nytimes.com/2022/01/05/ups…

maristpoll.marist.edu/polls/npr-mari…
Jan 3, 2022 • 4 tweets • 1 min read
It's probably foolish to think a NYC mayor will successfully translate into being a national political figure, but I still think Eric Adams would be in my top 5 for "who will be the next Democratic presidential nominee after Joe Biden?". He's going to be good at getting media attention and he has a chance of carving out a niche that's different from what other Democrats are offering. That's a potentially powerful combination given how primaries are conducted nowadays. Also, the competition isn't great.
Dec 30, 2021 • 4 tweets • 2 min read
You can squint and see the makings of cases leveling off in Manhattan (left-hand chart) but not really the rest of NYC (right-hand chart). Hard to know too much because of testing delays, though.
coronavirus.health.ny.gov/positive-tests… And this is probably best source for hospitalization data in NYC.

Good and bad news.

Bad news: they're growing fast, will keep growing.

Good news: *relative to the number of cases* lagged by a week, hospitalizations about 5-6x lower than last winter.

coronavirus.health.ny.gov/daily-hospital…
Dec 27, 2021 • 5 tweets • 2 min read
So, these results suggest that people who responded to this poll are taking something like 1.2 at-home COVID tests per 30 days, and 0.7 out-of-home COVID tests per 30 days. Not a representative group, but still interesting that at-home tests are the predominant method! By comparison, the average American is taking something like 0.1-0.2 out-of-home tests per 30 days. So, yeah, my followers like getting tested.

But we don't have a firm grip on how many *at-home* tests Americans are taking, which are largely going unrecorded.
Dec 19, 2021 • 4 tweets • 1 min read
Certainly, progressives were right that decoupling BBB and the infrastructure bill made it more likely that infrastructure would pass without BBB.

However, the alternative wasn't necessarily both passing. Perhaps more likely, neither would have passed. Lumping bills together tends to work better when you *have votes to spare* and can do some logrolling, i.e. Senator A doesn’t love Bill Z and Senator B doesn’t love Bill Y, but they’ll agree to vote for them in exchange for the other senator’s vote on the other bill.
Dec 15, 2021 • 4 tweets • 2 min read
Interesting that Republicans (64%) and Democrats (63%) are about equally likely to say they're "worn out" by lifestyle changes because of the pandemic (though Republicans are more likely to say that they're a *lot* warn out).

monmouth.edu/polling-instit… Needless to say, Democrats tend to be more supportive of restrictions, but also tend to face more restrictions in and/or are more likely to abide by them, which is why there's a lot of fatigue on both sides.
Dec 8, 2021 • 4 tweets • 1 min read
Not great that surveys are overestimating vaccine uptake (though, I'd note that the CDC's data has some issues too).

Both on vaccines and other COVID precautions, there are likely some fairly big biases in surveys toward people who are more conscientious/careful about COVID. Why?

1) People who are staying home b/c of COVID easier to reach
2) COVID concern tends to be higher among high-edu, high-news consuming, high-social-trust Americans, who are more likely to respond to polls
3) Maybe some social desirability bias (people exaggerate their caution)
Dec 6, 2021 • 8 tweets • 3 min read
To this good thread explaining why the "sentiment analysis" cited in the @milbank WaPo article this weekend is complete crap—the analysis was used to make the claim that the press is just negative toward Biden as Trump—I'll also add a couple of comments based on their data. 1/ These are the articles the algorithm found, out of more than 40K stories, were the most favorable toward Biden. It's just totally random. Lots of stories about the stock market. Many have nothing to do with Biden at all ("Haiti President Assassinated At Home, Wife Wounded"). 2/
Nov 3, 2021 • 4 tweets • 1 min read
I don't think this follows. In most recent elections, Democrats lost ground in rural areas but gained in the suburbs. Last night, they lost ground in the suburbs too. That's different and it's worth asking why (and, yes, the answers are probably complicated and multi-causal). This differs, by the way, from the California recall, where red areas got redder and blue areas got bluer or at least didn't shift much from 2018/2020. Of note, Elder was a Trumpist candidate while Youngkin and Ciattarelli were not exactly.