Nate Silver

Nate Silver @NateSilver538

US daily numbers via @COVID19Tracking: Newly reported deaths Today: 1,726 Yesterday: 1,510 One week ago (5/6): 1,949 Newly reported cases T: 22K Y: 20K 5/6: 25K Newly reported tests T: 336K Y: 289K 5/6: 242K Positive test rate T: 6% Y: 7% 5/6: 10% — PolitiTweet.org

Nate Silver @NateSilver538

If you look at *states* by the way, you get a pretty different impression. The top-surging ones in the report NBC & The Guardian highlighted were MN, DC, NH, WI, TN, NE, KS, MD, IL and IA ... not really much of a red/blue or urban/rural pattern there. — PolitiTweet.org

Nate Silver @NateSilver538

Let's try a test then: predict what the top 10 surge areas will be 2 weeks from now. (Must be specific cities/metro areas and you must define which metric you're using to define a surge.) — PolitiTweet.org

Nate Silver @NateSilver538

@JDVance1 @JeremyKonyndyk There's also difference between what public health folks actually say/think, and how the media characterizes the views of public health folks. — PolitiTweet.org

Nate Silver @NateSilver538

@willwilkinson "Rapidly" is doing a lot of work there. The evidence on the effects of a *gradual*, *cautious* reopening on transmission is a lot more uncertain, and that's in fact a better description of what most (indeed, almost all) states/countries are doing. — PolitiTweet.org

Nate Silver @NateSilver538

Indeed, folks need to be much more careful with the "Get ready for cases to EXPLODE in [state/county]!" types of headlines They tend not to recognize that behavioral and regulatory changes are often incremental, and the data is noisy—and they have often turned out to be wrong. — PolitiTweet.org

Nate Silver @NateSilver538

@jbarro tbh wouldn't surprise me if Biden wins the nomination. — PolitiTweet.org

Nate Silver @NateSilver538

@skepticalsports @MattGlassman312 Right but doesn't that just prove buffalo is a distinct flavor, which can be combined with other flavors as desired? — PolitiTweet.org

Nate Silver @NateSilver538

@MattGlassman312 Obviously true and not petty. — PolitiTweet.org

Nate Silver @NateSilver538

@aedwardslevy People forget that the War of 1812 actually happened in January — PolitiTweet.org

Nate Silver @NateSilver538

My value-add is not my expertise on public policy but in helping to provide guidance on what messy real-time data looks like when debated under fraught political circumstances. I think people should be alert to the possibility of noisy, confusing, possibly misleading signals. — PolitiTweet.org

Nate Silver @NateSilver538

3) Some factors are tending to produce improvement (mask-wearing, warmer weather, more testing) at the same time opening up may increase spread 4. On top of all that, the data is inherently fairly noisy and there are lots of debates about how to account for tests, etc — PolitiTweet.org

Nate Silver @NateSilver538

Yeah there are like 4 confounding issues: 1) "Opening" is often incremental (many things were never closed and only some of the closed things are opening) but is portrayed in the media as absolute 2) People's actual behavior is only loosely tied to a state's official policies — PolitiTweet.org

Nate Silver @NateSilver538

Folks, I don't have a prediction for what happens next. There are forces pulling in opposite directions (less distancing, but more masks, weather may matter, etc.) so the range of possibilities is wide. But the numbers HAVE gotten appreciably better over the past week or so. — PolitiTweet.org

Nate Silver @NateSilver538

Tuesdays are typically awful days as states work to clear weekend backlogs and can report a lot of cases and deaths. While 1,500 deaths in a day is indeed awful, it is better than the past 4 Tuesdays, where between 2,128 and 2,555 deaths were reported each time. — PolitiTweet.org

Nate Silver @NateSilver538

US daily numbers via @COVID19Tracking. Much better than past Tuesdays. Newly reported deaths Today: 1,510 Yesterday: 837 One week ago (5/5): 2,527 Newly reported cases T: 20K Y: 18K 5/5: 22K Newly reported tests T: 307K Y: 388K 5/5: 259K Positive test rate T: 7% Y: 5% 5/5: 9% — PolitiTweet.org

Nate Silver @NateSilver538

@IChotiner Yeah, though from September onward ~all polls will be LV polls so we'll be able to kind of ignore the issue from that point onward. — PolitiTweet.org

Nate Silver @NateSilver538

@IChotiner In theory, states where Democrats are relying on a younger, more diverse electorate (e.g. GA, TX) might show Trump gaining ground among likely voters whereas "white-working class" states (e.g. WI, MI) might show little gap or Biden gaining ground among LV's. — PolitiTweet.org

Nate Silver @NateSilver538

RT @gelliottmorris: @NateSilver538 FWIW and it's not not apples to apples since you're talking about change over time, but there's currentl… — PolitiTweet.org

Nate Silver @NateSilver538

@robbysoave Yeah, that was a bad use of polling. — PolitiTweet.org

Nate Silver @NateSilver538

So with a Republican incumbent this year and a Democratic challenger, those effects might roughly cancel out. FWIW, there also wasn't a big LV vs. RV spread in 2016 and 2018. — PolitiTweet.org

Nate Silver @NateSilver538

This is in the weeds a bit, but it's not clear that we should expect likely voter polls to be much different than registered voter polls. Historically, Republicans tend to gain 1-2 points on Democrats in LV polls. But challengers tend to gain 1-2 points relative to incumbents. — PolitiTweet.org

Nate Silver @NateSilver538

@mattyglesias It looks like people had to volunteer a "don't know" response instead of it being prompted in the question, which may overstate the degree of confidence but.... yeah. — PolitiTweet.org

Nate Silver @NateSilver538

I do wish CNN had provided a "don't know" answer on this question, because "there are a ton of unknowns so nobody can say with much confidence whether the worst is behind us" is probably the right answer. — PolitiTweet.org

Nate Silver @NateSilver538

So.... yeah, this is not surprising, but still interesting. It's also reflected on social media, where there are big partisan splits in how people tend to react to optimistic/pessimistic news. — PolitiTweet.org

Nate Silver @NateSilver538

Look for reporting and analysis that makes some effort to look at the data in a reasonably systematic or comprehensive way, acknowledges uncertainties and ambiguities, etc. This is not very common, unfortunately. — PolitiTweet.org

Nate Silver @NateSilver538

With 50 states, 100s of metro areas, and 1000s of counties, all of which have different reporting standards and quirks, and are doing different amounts of testing, etc., it's ludicrously easy to cherry-pick examples to support whatever point you might want to make. — PolitiTweet.org

Nate Silver @NateSilver538

@BenjySarlin It might actually be maximally relevant now. Early on, there was broad agreement. In the long run, people will likely just make whatever decisions are best for their personal circumstances. But right now, folks are getting mixed messages and partisan cues likely matter a lot. — PolitiTweet.org

Nate Silver @NateSilver538

Yeah, there should be close to zero doubt that Trump increases support for lockdowns among liberals and other people who don't like Trump. Also possible that he increases support for lockdowns *overall* since he's more unpopular than popular. — PolitiTweet.org

Nate Silver @NateSilver538

RT @geoffreyvs: Excellent article from @leedrutman on the vote-by-mail landscape and why mail ballots don't fundamentally advantage one par… — PolitiTweet.org