Meidas says it better than I could (initial monologue of this video).
Polls aren’t broken per se, but the way that polls used to be covered – so-and-so event has just happened, let’s see the effect on the polls – is archaic. It’s just not how things work now.
I don’t think this is indicative for November, but Trump was +27.6 points in final averaged South Carolina polling and +30.3 points in the election result.
More concerning to me, as a Democrat, is that Trump beat the final general election polling averages in both 2016 and 2020. Two elections is way too few to be sure Trump will beat the polls this coming November, but I sure wish I could say it went the other way.
For 2020, do you have a cite? I didn’t think that was the case.
Looking at some major poll aggregators, Trump did make up some ground in the popular vote between the late polls and the general election results. Look at the “Polling aggregation” section:
Nationwide opinion polling for the 2020 United States presidential election - Wikipedia
A lot of D’s, I think, are missing the point. It’s not about whether Biden will win or not (and there is a very real possibility he won’t beat Trump.) Rather, it’s also about how much he will win by.
The attitude of “Even if Biden just ekes it out by a hair, all’s well that ends well” neglects Congress. Even with just Mike Johnson and a tiny majority in the House, we can see the immense damage that MAGA Republicans can do with the slimmest of advantages. The Democrats need some strong coattails.
This isn’t the Super Bowl, where winning by 1 point is as good as winning by 30.
The only thing to do about it is to work as hard as we can for Biden and other Democrats. Barring an unexpected tragedy or massive new scandal, Biden will be our nominee. There’s no other scenario that gets another nominee. And those two scenarios would likely mean an even harder climb for Democrats.
Biden will be our nominee, and I’m fine with that. I don’t believe anyone else would be any more likely to win (or win bigger). He’s already done it once. Democrats just always worry. Abortion and IVF will be massive issues in the summer and fall, and Republicans still don’t have anything close to a credible answer that appeals to anyone but hardcore Christian conservatives. I think we’ll be okay.
Obama consistently polled above Romney, in the popular vote horse race, until September, when the race narrowed. So maybe you are remembering poor analysis.
But that’s not my main point.
My issue here is the idea that naysaying and negativity is bad for my side.
Suppose Biden describes himself — I believe accurately — as an underdog. Would this hurt him? Well, I think you could find social psychology research on this going both ways. But I will concede it being possible that self-describing as a political underdog is a mistake.
What I find utterly implausible is that calling Joe Biden an underdog on SDMB could hurt him. There is no reason to skew our play-analysis here with rose-colored glasses.
Agreed, although I doubt much can be done there.
Suppose Trump’s strategy of delaying criminal trials until after the election works for everything except the 34 felony counts of New York State business records falsification. Seems plausible to me. And suppose that he’s sentenced, late this summer, to a year of incarceration. And the judge won’t release him pending appeals he is unlikely to win. Sounds to me a Democratic wet dream – but not implausible. Donald then spends the last month or two, before the election, and election day, as a guest of the State of New York.
DJT would significantly drop in polls, resulting in GOP congressional candidates, if not exactly withdrawing their Trump endorsement, making grave faces when asked about it.
Result: Almost all voters, who normally ticket split, vote against Trump – and for the Republican House and Senate candidate.
Unless and until there’s a guilty verdict, polling this is impossible. But that’s my prediction of how the polls, and election results, would come. Not ideal, but a GOP congress could be an unstoppable railroad train.
Interesting findings from yesterday:
If polls are significantly overestimating his support in the primary, they may also be overestimating his support in the general. I’ve suggested this before, but perhaps Trump voters are more likely, right now, to talk to pollsters, than other voters.
The situation we have has not been there before. We never had a Trump before. I think there is a lot of hatred on one hand, but also a lot of disgust with politics in general. Many people have switched to a “no politics” mode in famlily and public situations. If asked to answer a poll, the immediate reaction is to hang up. There is almost no point in their minds, and my, to try to predict this and then to live in some panic state for months. I do that, but the disinterested people just turn it off in their minds. They have lives to live. Also, most people who do vote, realize the elctoral college situation is just unpredictable. Especially when both candidates run about 55% disapproval.
And that’s leaving out how so many polls are written in such a way that by answering, you’re giving those funding the poll a “Heads I win, tails you lose” support for their POV no matter what. By no means all of course, there are plenty of well-run polls out there that try for honest results, but there at least as many that load the language of their polls colder and harder than a stacked deck from a backroom cardshark.
I wouldn’t put too much stake in primary polls; clicking your link for South Carolina, there were only five polls of that race, none by pollsters I’ve ever heard of. It’s not surprising that the average of those polls might be pretty far off from the actual results. It’s mildly reassuring to see that the polling errors all seem to be running in the same direction, though.
I wonder how much stronger this reaction has become over the past few years. Phone habits are changing.
Everybody knows that young people only text, and never answer phone calls. But, as time marches on, those young people are getting older (and more likely to vote). And us old folks, who did answer the phone and talk to the pollsters in previous years, are also learning new habits. Smart phones are now standard…the old flip phones no longer in use, even in old-age homes. Every year, more and more old folks abandon their landline phone entirely. And presumably, more and more of us are irritated by, and suspicious of, unsolicited calls.
Lots of analysis has been done about the accuracy of polls taken 4 or 8 years ago. But I wonder if the conditions have changed enough to make that analysis less relevant for the current elections?
(and yeah, I see the irony…The only way to test this would be to take a poll. 
)
It’s risible to me that, although polling is a multimillion dollar industry which a lot of very smart people have devoted their professional lives to, people say things like “But I wonder if they’ve realized most people don’t answer their landlines anymore!”.
No question, as time goes on polling becomes more difficult,and it’s certainly not unreasonable to think that at some point in the future, technological and social changes may make polling unreliable. But it hasn’t happened yet; polling accuracy for the 2022 midterms was the best it’s been in several election cycles. There’s no reason to think it’s suddenly become completely unreliable in the last two years.
Agreed. You might like this Nate Silver link, although it’s partly behind a paywall:
The White House is betting the election on a theory of skewed polls
I think Silver’s word “unskewing” here is a polite way of saying — poll denialism. My takeaway: “banking on an error in your favor is usually a red flag.”
Oh, I’m a paid subscriber to Nate Silver’s blog! 
The thing is people only look at the last polls to see how accurate they are, nobody looks at polls from early March.
i received a text message this evening offering for Florida voters to take a poll. Figured what the hell.
Asked phone number, then party affiliation. Once i selected “Democratic” the poll was “Thank you, that’s all.”
I should’ve went with unaffiliated just to see how biased the questions were and maybe pretend my little blue opinion means a damn in this redneck Trumpopolis of a state.
It’s possible that it was an intentionally-biased poll; it’s also possible that it was a legitimate, neutral poll, which had proportional quotas for Democratic, Republican, and independent voters, and that by the time they had contacted you, they had already filled their quota for Democrats.
Nothing especially wrong with them, either. (That is, polls in September-October, e.g. in 2022).