Quoth Argentine Soccer Fan... "How realistic is it that all those polls could be wrong?" We had a massive weeks long discussion in 2016, and a very short discussion in 2018. Let's see what 2020 holds in store. Were the polls generally accurate, wildly inaccurate, or only deficient because they did not factor in a coup? To quote the eminent cartoon journalist Kent Brockman... "Only time. Will tell."
Another thing to keep in mind from 2016... https://www.realclearpolitics.com/epolls/2016/president/pa/pennsylvania_trump_vs_clinton-5633.html https://projects.fivethirtyeight.com/polls/president-general/pennsylvania/ A LOT more polling has been done on swing states in the final few days. We didn't just have big shifts due to Comey and other events. We had relatively few polls. We have more data. If there was a shift, we'd have seen it. Seriously. If the polls are wrong, we're not going to know the *causes* for weeks, if not months.
The extradimensional invasion was a bit of a twist but I can't shake the feeling that 538 should have had some kind of additional uncertainty factor to cover the possibility. How did they miss that? I, for one, welcome our new shoggoth overlords.
In case you haven’t noticed something interesting on RCP this time around: they seem to be cherry picking polls. They have completely excluded some high quality polls, they’ve adjusted their time window, and they’ve been flooded with partisan polls in the last few days. So they’re showing a horse race narrative in places like PA, but if you took a rilling average of the last 10 or so C+ (Rasmussen) or better polls, then leads are holding. Re: polling errors and their diagnosis. I think that if there are significant errors there’s a very good chance we’ll know what’s behind them within a few days. We knew 80% of the story last time by the end of the week. The likely voter models overestimated Black turnout and underestimated both white non-degreed and Hispanic turnout. We can look at turnout and margins at the county level vs county demographics by race, ethnicity, age, and education and pick that up quickly. It’s not something you’d chuck into an academic paper by Friday for review, but it’s likely to be 80% of the eventual story. The only real time issue is scraping/entering county totals. Nate Cohn seems to have surpassed Nate Silver in terms of understanding the statistical nuance at the local level. He also relies upon geospatial regression techniques rather than standard regression, which is extremely helpful when we’re talking about white working class Rust Belt v Sun Belt, etc. His tweets and Upshot will be a go to source for this thread.
Is this also the exit polling thread? Why we men should not be allowed to vote. 1323825846365396992 is not a valid tweet id
CNN How do you exit pol mail-in voters? I have a feeling exit polling is going to have a pro-republican bias.
Of course. Being that this is Matt Bruenig, it's hard to say whether this is a case of him not understanding that, or ignoring it in order to bolster his anti-Democratic Party, pro-Red/Brown alliance narrative.
I think in retrospect this was clearly correct: SCOOP: For months, senior Biden advisers focused on Hispanic and Black voters have raised concerns to leadership about insufficient investment. They say they have been ignored. Now, they are increasingly worried with just four days left. https://t.co/ndlb7QTpmH— Tyler Pager (@tylerpager) October 30, 2020 The other thing the polling missed was just how turnt white people are for Trump. Came out of the woodwork to vote for this man.
Apparently very hard to poll, I guess. I have a couple of Mexican-American friends who have some very spicy takes on the Mexican-American vote in the Rio Grand Valley, by the way. I am not going to write them down, but I imagine there will be some very interesting features written about these brand new voters in the weeks, months and years to come.
This will probably be a thread for "what internal polling got wrong" as well. That happened with Clinton in 2016 in PA: the campaign got reports from the field that were going against the rosy poll forecasts, and as a result, campaign workers got more and more anxious as election day approached. And the campaign workers who used to take their coffee breaks at my local coffeehouse got more and more panicked as they were more and more ignored, until it was too late.
If the pre-election polls were crap, we probably shouldn't put a whole lot of weight into the exit polls either. Seems to me that public opinion polling may simply be unreliable right now. If that's the case, then we shouldn't be relying on it for predictions, and it shouldn't be shaping our understanding of events.
Before we get to some solid analysis, there are going to be articles like this that raise some interesting questions, whether they have merit in the long run or not... For people who think Nate Silver should be homeless and 538 be declaring bankruptcy by the close of business.. https://slate.com/news-and-politics/2020/11/fivethirtyeight-upshot-polls-useless.html?via=taps_top . . . .At least someone was having fun. Never has this form of journalism’s close relationship to the analytics movement in sports media seemed more obvious, or more irritating. And never has the type of analysis they peddle felt more useless. “The narrative here fairly dumb overall,” FiveThirtyEight’s Nate Silver tweeted at 10:15 p.m. on election night, defending against accusations of a 2016-style polling miss as it became clear that Joe Biden was not racking up the early wins that Democrats had been hoping for. Silver’s tweet could serve as something of a mission statement for him and the other members of his coterie of data-driven political analysts: Pundits and traditional campaign journalists are in thrall to soft “narratives” about momentum and rely on anecdotes and conversations with voters. On the other hand, they believe, if you bloodlessly process a significant enough quantity of polling data, you can see the truth of what’s happening in the race. However the 2020 presidential election turns out in the end, it’s become clear that the polling analyses are themselves a “narrative,” one that can also obscure as much as it reveals. . . . . The polling gurus portray themselves as objective number-crunchers, unswayed by human bias or emotion. But in truth, they are in the reassurance business. Over the past weeks and months, after any troubling piece of news came out about lagging minority turnout or legal challenges to mail-in voting, Democrats could check FiveThirtyEight and see Biden’s odds at 89 out of 100 (or, if they were really feeling glum, check the Economist’s G. Elliott Morris and see them at 19 out of 20) and feel like things were still under control.* But they’re hawking a false sense of certainty—and, presumably, racking up earth-shattering levels of web traffic in the process.
Or to put it more succinctly: garbage in, garbage out. 538 can crunch all the numbers they want, but if the numbers are inaccurate to begin with, what's the point?
All for dunking on these guys today b/c I don't like the way they present their findings but at the same time, they absolutely nailed the midterms. It just seems like the turnout of low-propensity voters broke the models.