Nate Silver

Nate Silver @NateSilver538

Like, we desperately need to center the conversation on questions like: What are viable approaches to keep R <=1 for the next ~12 months? What are the costs/benefits of each approach? Are we willing and able to do any of them? If not ... what does mitigation look like? — PolitiTweet.org

Nate Silver @NateSilver538

Maybe part of a theme where there's a widening disconnect between people's outwardly-expressed views ("political positions") on what do to about COVID and their revealed preferences, both as demonstrated through their own behavior and through what their governments are doing. — PolitiTweet.org

Nate Silver @NateSilver538

An interesting gap here where people think "stay home" is still the best advice, and that hasn't changed very much. But, when you ask people how they're behaving personally, people are actually doing quite a bit less distancing over the past ~3-4 weeks. https://t.co/Sqsg2mfKkT https://t.co/zMWb5n9WUF — PolitiTweet.org

Nate Silver @NateSilver538

RT @perrybaconjr: We looked at how the Biden VP is process is fraught with people kind of hiding their real motives. https://t.co/HuD8YiUBIW — PolitiTweet.org

Nate Silver @NateSilver538

I would note that even if it's true, it's likely a bit of a quirk that 4 of the past 5 elections (all bit 2008) have been *so* close. (A winning margin of <4 points.) A lot of midterm elections have been landslides during this time frame. — PolitiTweet.org

Nate Silver @NateSilver538

It's one of those things where if it's true, it's pretty consequential (might mean that winning margins are only ~half as high today as the historical average), but it's not entirely clear that it's true (data is noisy)! — PolitiTweet.org

Nate Silver @NateSilver538

The idea that presidential elections are getting closer because of greater polarization is highly plausible but not proven. This chart shows polarization in Congress (via DW-Nominate) vs. the popular vote margin for the winning candidate in presidential elections since 1880. https://t.co/42wwhToq3U — PolitiTweet.org

Nate Silver @NateSilver538

RT @ClareMalone: This week's podcast was a primer on today's primaries https://t.co/nVP84L1wWs — PolitiTweet.org

Nate Silver @NateSilver538

One implication of the US experience, in which the states having bad outbreaks now are almost literally the opposite of the ones that were having bad ones in March/April, is there's probably a bit too much chicken-counting in terms of which *countries* have done well with COVID. — PolitiTweet.org

Nate Silver @NateSilver538

@JADubin5 @thecity2 @COVID19Tracking That is probably much of it. There could also be lags, though (so we'll see what the numbers look like in ~10 days). Also, possible that treatments are improving and/or we're doing a less terrible job protecting nursing homes. — PolitiTweet.org

Nate Silver @NateSilver538

In positive news, Google searches for masks have ticked up after a long decline. And guess where there's been the biggest increase over the past 7 days? Arizona, which has had (perhaps) the worst recent spread and where Phoenix just issued a mask mandate. https://t.co/OovS8hEwNS https://t.co/MTeVRvLu7N — PolitiTweet.org

Nate Silver @NateSilver538

Par for the recent course. Deaths down week-over-week. But the positive test rate and the number of cases has risen quite a bit. — PolitiTweet.org

Nate Silver @NateSilver538

US daily numbers via @COVID19Tracking: Newly reported deaths Today: 285 Yesterday: 297 One week ago (6/15): 375 Newly reported cases T: 28K Y: 26K 6/15: 19K Newly reported tests T: 475K Y: 499K 6/15: 448K Positive test rate T: 5.9% Y: 5.3% 6/15: 4.2% — PolitiTweet.org

Nate Silver @NateSilver538

@neeratanden And the fact that people are completely numb to all of this now is at least suggestive of the fact that a little bit more nuance and measuredness would have helped. — PolitiTweet.org

Nate Silver @NateSilver538

@neeratanden Really? I think having said back in April that the re-openings weren't likely to cause big immediate spikes because the re-openings were pretty measured, but that this could lead to complacency as the re-openings became broader later on, has to count as a pretty good prediction. — PolitiTweet.org

Nate Silver @NateSilver538

So, this is a welcome (and overdue) change in tone from Gov. Abbott after COVID increases in Texas. Trying to de-polarize the issue of mask-wearing would be one of the most helpful things right now. — PolitiTweet.org

Nate Silver @NateSilver538

Like the way this whole thing has played out has been regrettably predictable. People would make vastly over-reachi… https://t.co/DNHRV1tkQa — PolitiTweet.org

Nate Silver @NateSilver538

Relatedly, I don't think there should be a lot of triumphalism over "see, we said cases would spike in [insert state] once it reopened" as those predictions often implied an IMMEDIATE spike in cases within 2-3 weeks of re-openings even though openings were initially very limited. — PolitiTweet.org

Nate Silver @NateSilver538

The world is complicated. * Very likely that relaxation of distancing is contributing to COVID cases increasing * But there are many states/countries that where distancing is decreasing but cases are not increasing * Formal "opening" status is only loosely related to distancing — PolitiTweet.org

Nate Silver @NateSilver538

@nataliexdean Yeah, maybe looking at behaviors that directly get at nondistancing, e.g. restaurant reservations (although yes you can eat outside), movie theaters, might be better. I know some folks in NYC think that restaurant data and data on public transit use is at least reasonably useful. — PolitiTweet.org

Nate Silver @NateSilver538

@nataliexdean One in particular: I wonder if mask wearing is much higher in urban areas and therefore has started to neutralize a lot of the effects from density that we saw earlier. — PolitiTweet.org

Nate Silver @NateSilver538

@nataliexdean In general, both re-opening data and mobility-tracking data has been only weakly predictive both when I've noodled around with my own analysis and when I've looked at what others have done. But of course we know there must be SOME effects, so yeah, confounders. — PolitiTweet.org

Nate Silver @NateSilver538

@nataliexdean For sure but FWIW when you try to control for a lot of the confounders these findings remain pretty robust. Would be better if looking at county-level data, etc. — PolitiTweet.org

Nate Silver @NateSilver538

RT @ABCPolitics: Can Trump ride a midwest path to victory? “Trump can absolutely win reelection but he definitely has his work cut out for… — PolitiTweet.org

Nate Silver @NateSilver538

RT @maggiekb1: Sometimes @FiveThirtyEight, we go looking for data about a problem and end up finding an entirely different story ... becaus… — PolitiTweet.org

Nate Silver @NateSilver538

@gelliottmorris The movement data tends to be only weakly (if at all) predictive. Mask-wearing might be interesting, though. — PolitiTweet.org

Nate Silver @NateSilver538

@gelliottmorris Yeah I'm using https://t.co/LwOqS5Fl1c. Spread is higher in places where it's likely that more people are indoors. It's also higher in places that didn't get hit hard before (likely based on a combo of some people now having immunity and perhaps people being more careful). — PolitiTweet.org

Nate Silver @NateSilver538

I would not be surprised if a lot of the early research on weather conflated temperature effects vs. indoor/outdoor effects. If you're comparing places in March, then people will be outdoors more often in warmer places. But that's not necessarily true in the summer. — PolitiTweet.org

Nate Silver @NateSilver538

What's also interesting is that if you add temperature to this, higher temps are associated with lower spread (as most research predicts), controlling for other factors. So heat helps, but being outdoors helps, and you don't want it to be so hot that folks do everything inside. https://t.co/IQY3h8Dg24 — PolitiTweet.org

Nate Silver @NateSilver538

Saw someone suggest "cooling degree days" (basically the number of days that it's hot enough out that you'd want the AC on) as a proxy for states where it's currently too hot for people to want to be outdoors. It is indeed quite predictive of current COVID spread. https://t.co/4iooEtXR7N — PolitiTweet.org