My biases are on the table -- I'm a Democrat -- but trying to be objectively analytical as possible, I have to think the last 24 hours have been extremely favorable to Biden's re-election. Here's why 1) If anyone had any questions about whether Biden had the energy for a second term run, those questions have been answered. There wasn't anyone credible running against him in a primary, but a weak performance could have had ambitious candidates thinking about it. Not anymore.

Parsing the various 2024 GOP candidates strategies, and here's what I got (dissenting views welcome): 1) Trump: Demonstrate dominance in all cases, as if he were some alpha gorilla in a nature documentary rather than a candidate for office. See: him talking about DeSantis begging for his support. Seems to believe is he acts dominant, voters will gravitate toward him

Many liberals (me included) are fond of this quote: “Conservatism consists of exactly one proposition, to wit: There must be in-groups whom the law protects but does not bind, alongside out-groups whom the law binds but does not protect.” slate.com/business/2022/… I'm curious, though: what similarly pithy summary would conservatives offer of liberalism? Or for that matter that liberals should liberals offer liberalism?

40 minutes and counting. At about 30 minutes they said "we have another 8 minute survey"

That's all one survey, saying you have another survey that's clearly related to the first is just a farce. C'mon people.

https://twitter.com/kwcollins/status/1622974244383784960

If you're fielding a survey this long, prepared for respondents to be doing something else while they answer.

Finished Damascus Station by @mccloskeybooks last night, and wow, it's the best spy novel I've read in a long time. If you liked Red Sparrow, with the emphasis on process, this one is for you. Credit to @ThatWasChatter for first turning me on to it (though that podcast was, over a year ago, so shame on me for not actually reading it sooner). And pour one out for the dearly departed Bombshell podcast where I used to get my spy novel recs from @charlie_simpson

This is a good @mattyglesias piece but I think he mostly overlooks one of the greatest buses of historic preservation: classification of whole neighborhoods as historic, including buildings that are not themselves historic but are just sort of nearby slowboring.com/p/commemorate-… I live in the Takoma DC neighborhood, near but not in the Takoma Park (DC) historic district (the name is confusing). planning.dc.gov/publication/ta…

One problem with @Nate_Cohn's theory that Dems underperformed in NY relative to PA because of crime played a bigger role in NY: About the same number said crime was an important factor (85 and 87% respectively), though in PA, Dem won a majority of those votes while in NY GOP did This is from Votecast, btw. Other issues:

Abortion: NY 72%, PA 70%
Future of Democracy: NY 86%, PA 86%
Inflation: NY 93%, PA 93%

This @Nate_Cohn hypothesis of divergent state results is an interesting theory but I see some data points that cut against it, would love to get his take on them nytimes.com/2022/11/11/bri… (thread) First, it makes the claim that NY gov race was about crime not abortion. Systematic data would be helpful here, but my sense was that Hochul did a lot of ads on abortion.

It's not a perfect analogy, but I've been thinking a lot about how current polling market resembles Gresham's Law of "Bad money drives out good" en.wikipedia.org/wiki/Gresham%2…

https://twitter.com/KSoltisAnderson/status/1590897660525178880

There were a lot of cheap and poorly done polls this cycle, and fewer (but still quite a number) of well done, more expensive polls. Those two were taken at equal value by commentators, and by some of the averaging sites.

First, weather forecasters have exponentially more data points to operate from than do election modelers. There's weather every day that can be used to refine and build models, there are US federal elections every 2 years.

Second, polls actually looked pretty good this year.

https://twitter.com/samsanders/status/1590420541063979008

Third, polling and election modeling based on polling -- while certainly not as precise as still beats the alternatives. It looks better this year than fundamentals-based predictions. It is a lot better than n=1 person-on-the-street interviews.

This Twitter Spaces with @elonmusk is interesting, but also highlights the inherent contradiction in the Twitter Blue strategy they're pursuing. He highlights widespread verification as a strategy for reducing bots (good!) but trying to monetize that verification undercuts it I also like the idea of prioritizing tweets or notifications from verified users. But if you're charging for that verification, you're asking people to pay to be heard, rather than pay to listen, literally the inverse of traditional media paywalls.

While you're waiting for election results to be finalized so you can evaluate the polls (you are waiting, right?), flagging some very cool recent-ish articles from @JSurvStatMeth you might have missed. (Nerdy thread) 1) "Exploring the Feasibility of Recruiting Respondents and Collecting Web Data via Smartphone: A Case Study of Text-To-Web Recruitment for a General Population Survey in Germany" academic.oup.com/jssam/article/…

Some nuggets from the AP VoteCast (aka Fox News Voter Analysis) data. First up: there's no gender gap among urban whites People who said Trump *was* a factor in their vote were far more likely to support Democrats than people who said Trump was *not* a factor

As an outsider looking in, I'm very curious how the dynamics of the GOP will play out over the next few weeks. Specifically

1) Will DeSantis (or allies) make a head-on attack on Trump as a loser (I suppose some of this depends on how AZ plays out), and how will Trump respond? (I would not be entirely surprised if Trump, who has no party loyalty or really loyalty to anyone, is willing to blow everything up just to take down people who he things slighted him)

However the chamber control turns out, Democrats *massively* overperformed fundamentals-based expectations grounded in prez approval or inflation

We can't yet calculate polling error where we don't have final results, but we can say polls did better than pure fundamentals Based on simple correlations with historical data, if you based predictions solely on those factors, you'd expect 30-40 seats lost.

Source: presidency.ucsb.edu/analyses/the-2…

The @Yascha_Mounk article has gotten mocked in my (very COVID-cautious) corner of twitter. But I think it's worth it's unpacking to see how the arguments in favor of throwing all caution to the wind are made, and how they employ rhetoric to avoid uncomfortable facts. That article is here, for those following along: theatlantic.com/ideas/archive/…

Don't Look Up posits a world in which greed of US tech companies and media in attention are the main obstacles to preventing a global catastrophe, even though international cooperation was present. This seems to me to get the situation with climate change almost exactly backwards While there's no doubt that companies in extractive industries have been obstacles, others want to sell green tech. And while morning cable news shows may not help much with long term thinking, they're not the whole of media.

Would love see analysis of correlates of individual change (both to the left, and the right) using panel data like these

https://twitter.com/PollsAndVotes/status/1490731265116487680

https://twitter.com/Nate_Cohn/status/1490733201387667457

What kind of weak or leaning Democratic voters in 2016 became, by 2020, strong Democrats? Vice versa? Same for Republicans? Who became independent, and who affiliated?

I'd really like to see more open end follow ups on approval questions. Like asking people whether they approve or disapprove Biden's handling of COVID, and then asking "when they think about Biden's handling of COVID, in a few words, what are you thinking about?" Or after approval on the economy, ask people what aspects of the economy come to mind when they offer that evaluation

Neil Gorsuch is in the row behind us on this flight. Tempted to say, “Excuse me, I think you’re in Merrick Garland’s seat” This is a joke, I’m far too polite to ever harass someone on an airplane. But not too polite to think it

Exit polls should be used, if at all, to understand the election result, not to predict it. To do so, they are weighted to the election results and final turnout numbers. You can't do that weighting properly without results or turnout numbers.

So ignore early exit polls. If things are great for your team, or if they're terrible, you'll know in somewhere between 2 and 12 hours from now. So, just wait.