Nate Silver

Nate Silver @NateSilver538

In NYC, cases are down 11% and the positive test rate is down 10% in the past week as compared to the week before that. In the NYC burbs, cases are down 3% though the positive test rate is up 2%. Upstate, though, cases are up 11% and the positive test rate is up 17%. — PolitiTweet.org

Nate Silver @NateSilver538

It's a bit hard to know what to make of New York's COVID data since cases and positive test rates are now very low and I don't know how much it's worth focusing on small shifts. But FWIW, there are slightly different patterns upstate vs. downstate. https://t.co/1w2VrzC4FG — PolitiTweet.org

Nate Silver @NateSilver538

So all these positive tests among NBA players and other athletes may tell you something about where the disease is most prevalent now. Athletes are young and often have homes in current virus hotspots such as Florida and California. Likely a high number of social contacts, also. — PolitiTweet.org

Nate Silver @NateSilver538

A ~5% positivity rate among NBA players. That's fairly high; nationally, the positive test rate is ~7%, but it's presumably mostly people with symptoms or exposure getting tested and not a sample of the entire population. — PolitiTweet.org

Nate Silver @NateSilver538

@mattyglesias It certainly may be politically relevant, although the states with the biggest problems right now are really purple and not red. (If the takes were based on who is favored to in *2020*, it would look different). But these takes are not very rigorous as analysis of COVID itself. — PolitiTweet.org

Nate Silver @NateSilver538

Also, one of the lessons of the past few weeks might be: be careful with taking premature victory laps. We don't know where the NEXT spikes are going to come. Until we've made it through (at least) a year of this thing, avoid hasty conclusions on which states did best. — PolitiTweet.org

Nate Silver @NateSilver538

Here's a graph comparing current Rt to Trump's margin of victory in 2016. See a relationship? There is one, but it's weak and disappears once you account for weather and how many people had COVID before (the virus is spreading faster in places that weren't hit hard before). https://t.co/SdXe4IcgHZ — PolitiTweet.org

Nate Silver @NateSilver538

The relationship between the current rise in COVID cases and the partisanship of a state/city/county is much weaker and more complicated than the commentariat seems to assume. — PolitiTweet.org

Nate Silver @NateSilver538

Overall, the bevy of polls today are pretty consistent with what you'd expect given a 9-10 point Biden lead nationally. But the big news is that MI, WI and PA are starting to poll closely in line with the national averages, instead of being R-leaning as in 2016. — PolitiTweet.org

Nate Silver @NateSilver538

So our averages are not *quite* as optimistic as Fox on Biden in the South. But keep in mind that many polls now (like these Fox News polls) were conducted among registered voters. Our averages make a likely voter adjustment to those, which helps Trump by ~1 point or so. — PolitiTweet.org

Nate Silver @NateSilver538

Our averages in these states: FL:Biden +7.2 GA: Biden +1.4 NC: Biden +2.9 TX: Trump +0.2 https://t.co/nXOH1SNdLP — PolitiTweet.org

Nate Silver @NateSilver538

So, then, the real headline is the record number of cases, more than 40,000. 😬 There were a record number of tests reported also, so the positive test rate wasn't quite as bad as yesterday. Still, it's increased quite a lot from where it was a week or two ago. — PolitiTweet.org

Nate Silver @NateSilver538

My understanding is that in cases like these, where states report big one-off numbers because of changes in reporting standards, @COVID19Tracking may sometimes go back and try to figure out when the deaths *actually* occurred. But, it's hard. The quality of state data varies. — PolitiTweet.org

Nate Silver @NateSilver538

† New Jersey today decided to begin including *probable* deaths in its tally, which added 1,854 deaths to its count. Many occurred some time ago, so it's worth thinking of them in a separate bucket. Absent those, there were 646 deaths reported today. https://t.co/It7Lt0Xkfx — PolitiTweet.org

Nate Silver @NateSilver538

US daily numbers via @COVID19Tracking: Newly reported deaths Today: 646 + 1854† Yesterday: 722 One week ago (6/18): 693 Newly reported cases T: 42K* Y: 39K 6/18: 27K Newly reported tests T: 640K* Y: 512K 6/18: 518K Positive test rate T: 6.5% Y: 7.5% 6/18: 5.3% * record — PolitiTweet.org

Nate Silver @NateSilver538

My understanding is that in cases like these, when states report big 1-off numbers because of a reporting change, @COVID19Tracking will sometimes later go back and try to figure out when those deaths actually occurred. But, this is not always easy/possible; states' data is messy. — PolitiTweet.org

Nate Silver @NateSilver538

† New Jersey has now reporting 1,854 *probable* deaths, many of which occurred some time ago, as part of its tally. It's probably worth separating those out for now. Excluding those, there were 646 deaths reported today. https://t.co/It7Lt0Xkfx — PolitiTweet.org

Nate Silver @NateSilver538

US daily numbers via @COVID19Tracking: Newly reported deaths Today: 646 +1854† Yesterday: 722 One week ago (6/18): 693 Newly reported cases T: 42K* Y: 39K 6/18: 27K Newly reported tests T: 640K* Y: 512K 6/18: 518K Positive test rate T: 6.5% Y: 7.5% 6/18: 5.3% * Record — PolitiTweet.org

Nate Silver @NateSilver538

Notes: in the regression above, I'm specifying mask policy on a 4-point scale based on the chart in the first tweet. If you take the results at face value (which, maybe you shouldn't!) a strict mask policy would reduce R by around ~15% vs. no mask policy. — PolitiTweet.org

Nate Silver @NateSilver538

It may not be causal; mask policy may be a proxy for lots of other decisions that states take, i.e. their overall level of caution. Still, this is a stronger relationship than a lot of stuff I've looked at (mobility, restaurant reservations, partisanship, etc.) — PolitiTweet.org

Nate Silver @NateSilver538

If you look at current R's and control for weather and the total number of people infected as of ~2 weeks ago (states that had bad outbreaks before are having fewer problems now), then you do see some apparent effects of mask policies. https://t.co/GeZt7UoUd2 — PolitiTweet.org

Nate Silver @NateSilver538

A lot of confounding variables, but there may actually be something here. — PolitiTweet.org

Nate Silver @NateSilver538

RT @sfrostenson: One thing that sticks out to me in our latest polling average update is what we've seen for a few weeks now, polls point t… — PolitiTweet.org

Nate Silver @NateSilver538

RT @daveweigel: This was a poll of New York state, which Clinton won by 22 points. https://t.co/wxS8XhNOt6 — PolitiTweet.org

Nate Silver @NateSilver538

RT @jazzmyth: I made a visual explainer to accompany @laurabronner's @FiveThirtyEight article showing how flawed logic leads to a misunders… — PolitiTweet.org

Nate Silver @NateSilver538

Three Pennsylvania polls out today, all good for Biden, that have closed the gap a bit between the Electoral College & the popular vote. PA, where Biden leads by 8.1, is the current tipping-point state. We have him up by 9.5 points nationally. https://t.co/wKbW02AGO5 https://t.co/9dCiRkF0GL — PolitiTweet.org

Nate Silver @NateSilver538

Didn't realize it was this rare for a candidate to be >=50% at this point in the election. It worked out pretty well for Nixon in '72 and Reagan in '84. — PolitiTweet.org

Nate Silver @NateSilver538

Weird polls somewhat aside, Michigan and Wisconsin are probably the states where Trump's decline in the polls is most obvious and emphatic. https://t.co/t9jgEHjull https://t.co/o5UpDNgQ3Z — PolitiTweet.org

Nate Silver @NateSilver538

Here are the last 7 polls of Michigan: Biden +18 Biden +11 Biden +1 Biden +2 Biden +13 Biden +16 Biden +15 ¯\_(ツ)_/¯ — PolitiTweet.org

Nate Silver @NateSilver538

RT @laurabronner: We have evidence of racial discrepancies in police use of force. But some studies miss the extent of the problem by using… — PolitiTweet.org