Nate Silver

Nate Silver @NateSilver538

If Clinton had won in 16, you wouldn't have as many Trafalgars publishing Trump +2 in every swing state. Pollsters wouldn't be doing as much education-weighting. Etc. Biden would likely have a larger lead *in the averages* even if his true standing in the electorate was the same. — PolitiTweet.org

Nate Silver @NateSilver538

Market participants don't really get this IMO, so the markets sometimes wind up doing a sort of double-counting. They'll say "polls were skewed against Trump in 2016, so they might be again in 2020". Not at all crazy! But the polling averages *may already be accounting for this*. — PolitiTweet.org

Nate Silver @NateSilver538

This also helps to explain why prediction markets have been largely unable to beat publicly-available polling-based models (e.g. 538's) over the years. The polling averages already account for many of the desirable characteristics of markets. — PolitiTweet.org

Nate Silver @NateSilver538

People neglect the degree to which polling averages are themselves sort of a market-driven consensus. Their power of these averages comes as much from the fact that they incorporate different views about the electorate as that they provide for a larger sample size. — PolitiTweet.org

Nate Silver @NateSilver538

There are a handful of polling firms (e.g. Trafalgar and Rasmussen) that rather explicitly operate on the premise other polls are skewed against Trump. Including them in your polling averages (as 538 does) means the averages *are* accounting for a smidgen of a "shy Trump" effect. — PolitiTweet.org

Nate Silver @NateSilver538

This is not really about prediction per se but about attribution. If, in 2044, some data journalist comes along and says "welp, in 2020, the economy went into a recession, and yet again the incumbent lost!" that will be a pretty absurd description of why Trump lost (if he loses). — PolitiTweet.org

Nate Silver @NateSilver538

Given rapid improvement in many economic indicators, and fairly high approval for Trump on the economy, the idea that Trump is doomed to lose because of the economy seems pretty miscast at this point. It really might be a *comparative* strength of his. https://t.co/BvgKvmTa3x https://t.co/kHR5xi7tPr — PolitiTweet.org

Nate Silver @NateSilver538

RT @ashishkjha: I neither understand these projections nor do I buy it. To be clear, there are many hard days ahead. Many more Americans w… — PolitiTweet.org

Nate Silver @NateSilver538

@jonathanchait Basically, the model docked Biden a smidge when it looked like he wasn't getting much of a convention bounce, but now that Trump hasn't gotten much of one either, it's come full circle. The big sell-off in the stock market today also hurt Trump a bit. — PolitiTweet.org

Nate Silver @NateSilver538

@jonathanchait 1) If the polls remain stable Biden's probability will improve, but not necessarily steadily; would be mostly concentrated in the final few weeks. 2) Biden's gradually been gaining over the past several days as it's become clear that Trump wasn't getting much of a bounce. — PolitiTweet.org

Nate Silver @NateSilver538

FWIW, the poll train is probably going to slow back down at this point. I'm sure there will be some odds and ends, but we're not aware of too much high-quality stuff that's imminent, and pollsters that missed this "round" will likely mostly wait until after Labor Day. — PolitiTweet.org

Nate Silver @NateSilver538

Not many changes in our *forecast*, which has been stable and pretty unremarkable, but Biden has climbed back up to 71 percent—the same number he was at when we officially launched last month—after a brief jog down to 66/67. https://t.co/4qLlnZAfWT — PolitiTweet.org

Nate Silver @NateSilver538

With these polls, FL and PA flip places in our polling averages, with Biden going from +3.4 to +4.1 in PA but from +4.1 to +3.3 in FL. (The FL poll is a bit worse than it looks for Biden because previous Quinnipiac polls in FL had been better for him.) https://t.co/wKbW02AGO5 https://t.co/JQIORI9MI4 — PolitiTweet.org

Nate Silver @NateSilver538

Overall, the Fox + Monmouth + Quinnipiac state polls would seem to be quite consistent with Biden holding a 7-8 point national lead but Trump polling at more of a 4-6 point deficit in the tipping point states. — PolitiTweet.org

Nate Silver @NateSilver538

Quinnipiac has Biden +3 in Florida and +8 in Pennsylvania. Likely voters. https://t.co/N0yNnvsVxA — PolitiTweet.org

Nate Silver @NateSilver538

* Not necessarily in the sense of anything untoward, but in how you choose the group, how you pick the questions, how you steer the conversation, and how you summarize the results, there are basically infinite degrees of freedom to get the focus group to "say" whatever you want. — PolitiTweet.org

Nate Silver @NateSilver538

Focus groups are not good substitutes for polls, but I can imagine times when they provide useful *qualitative* insights. But it's not a coincidence that, in practice, they almost always wind up endorsing the prevailing media narrative about the race. Very easy to manipulate*. — PolitiTweet.org

Nate Silver @NateSilver538

So... a Fox News poll of 800 people yesterday showed Biden doing quite well in Wisconsin post-Kenosha. Not sure, therefore, why a non-randomly selected focus group of 10 people (8 of whom voted for Trump last time!) is really newsworthy at all. https://t.co/ws088l2PhI https://t.co/Ejw60HNypl — PolitiTweet.org

Nate Silver @NateSilver538

RT @FiveThirtyEight: Biden gets good polls in Arizona and Wisconsin — and a bad one in Pennsylvania. https://t.co/PDCOBZT9NU — PolitiTweet.org

Nate Silver @NateSilver538

Plan A: Hold the Midwest (WI/MI/MN/PA). Plan B: Lose some of those Midwestern states but win Florida. Plan C: Lose *some* of those Midwestern states and lose FL but win NC and/or AZ. Plan D: Get swept in the Midwest but win FL + NC + AZ. — PolitiTweet.org

Nate Silver @NateSilver538

North Carolina is the 7th-most likely tipping-point state after PA, FL [** big gap **] WI, MI, MN and AZ. It's part of what you might call Biden's Plan C. https://t.co/ajG88SznSA https://t.co/l9Ci1JdF5n — PolitiTweet.org

Nate Silver @NateSilver538

Biden +2 in North Carolina per Monmouth, which is a bit lower than Fox News but closely matches our average there (Biden +1.6). https://t.co/kNlJ1W6cTa — PolitiTweet.org

Nate Silver @NateSilver538

Some signs of increasing COVID cases in the Midwest, which is a region you might expect to have some issues ... didn't get hit hard before, lots of big colleges & universities, etc. https://t.co/pXQZgBOZUO https://t.co/CVAbnPQQp1 — PolitiTweet.org

Nate Silver @NateSilver538

@AdamSerwer @katherinemiller A lot of transplants. I'm still not sure pollsters are doing a great job of capturing a representative sample of Hispanics, especially Spanish-speaking ones. Can imagine the Native American population is hard to poll, too. — PolitiTweet.org

Nate Silver @NateSilver538

This is plausible! If Democrats feel like life as they know it is over if Trump wins re-election, they're probably not eager to lay additional money down on Biden and instead would want to hedge. — PolitiTweet.org

Nate Silver @NateSilver538

Biden is up *8* points in a *Fox News* poll of *Wisconsin* conducted *immediately after the RNC* and prediction markets have the race as a toss-up. — PolitiTweet.org

Nate Silver @NateSilver538

And... Arizona's is now the tipping-point state in our polling average. Biden's lead there (+4.7) is slightly larger than in PA (+4.3) or FL(+4.1). WI (+6.9), MI (+6.5) and MN (+6.0) are polling ahead of the tipping point, NC (+1.6) behind it. https://t.co/nWUo9X5F4y https://t.co/T8qa9B28E8 — PolitiTweet.org

Nate Silver @NateSilver538

There are also registered-voter versions of the the Fox News polls that show about the same results, with Biden +10 in AZ, +4 in NC, +8 in WI. Not great for Trump since he might hope to gain from the likely-voter screen. Loosely speaking, means no enthusiasm gap after the RNC. — PolitiTweet.org

Nate Silver @NateSilver538

The Arizona polls have become quite bifurcated, with a group of polls showing a toss-up and another group showing Biden ahead by margins ranging from 7-10, with little in between. Not always an easy state to poll. https://t.co/AniJJXutqh — PolitiTweet.org

Nate Silver @NateSilver538

They're also likely voter polls with pretty decent sample sizes. I'd note that Fox has generally had good numbers for Biden this year, but these generally reflect some improvement for him. They previously had Biden +4 in AZ, +9 in WI and +2 in NC, all among *registered* voters. — PolitiTweet.org