IIRC, Nate’s point was that once you do take the bias into account, one can use Rassmussen. And I do remember that he used Rasmussen with a very significant correction added to that pollster to every poll they made. Then the accuracy was close to reality…
Silver might be kind of off on an island on that one (though GIGObuster’s just-now post perhaps redeems Silver a bit). The Washington Post article you shared explains plainly why Rasmussen is getting shunned in the community of pollsters:
For years, Rasmussen’s results have been more favorable for Republican candidates and issues. During the Trump administration, though, the site’s public presence became more overtly partisan, with tracking polls sponsored by conservative authors and causes and a social media presence that embraced false claims that spread widely on the right. At times, Rasmussen’s polls actively promoted those debunked claims, including ones centered on voter fraud.
[In March 2023], for example, Rasmussen released data purporting to show that Republican Senate candidate Kari Lake (R) had won her gubernatorial election in November 2022. The route it took to get to that determination was circuitous and, to put it mildly, atypical. On behalf of the group College Republicans United, Rasmussen asked Arizona voters who they voted for in Lake’s race and, after weighting the results to exit polls — which is unusual — declared that, contrary to the certified tally, Kari Lake had won her race by eight points.
An election of 2.5 million voters is a better indicator of an election outcome than a retrospective question offered to 1,000 Arizonans four months later from a Republican-leaning pollster that is adjusting its results to a metric, exit polls, that is itself weighted to the election results. But Rasmussen trumpeted this revisionist look at the race loudly — including on Stephen K. Bannon’s podcast — as did Trump allies.
…
[In late 2023], Rasmussen again published dubious poll results on behalf of a right-wing organization. This time, the findings alleged to have uncovered rampant fraud in 2020, including that 1 in 12 Americans had been offered “pay” or a “reward” for their vote. Trump and his allies celebrated the poll; again, the results do not comport with the reality of there being no demonstrable wide-scale vote-buying scheme at the state or national level.
RCP brought Rasmussen back in — the latter coming up with a +8 point Trump outlier this week.
The polls are suggesting a huge shift in the electorate. Are they right?
Makes sense for Pennsylvania because, excepting college grads, the best-paying jobs for strong young people are in fracking.
But that won’t apply to other swing states.
The paragraph quoted above seems written to promote the horse-race narrative. “X, but Y! Y, but Z! It’s anybody’s hame, so pay attention to us! Whatever you do, give us your clicks and eyeballs!”
… and just like that: RealClear Polling (a) knocked all polls over two weeks old out of their average, which ditched a Trump +7 [Selzer] and (b) brought in a new Biden +3 poll (Issues & Insights/TIPP). That makes the Biden-Trump head-to-head closer in the RCP average than they’ve been in several months – 0.3% (yes … three-tenths of a percent).
Also: This is the first time in a long while that Biden has led Trump in more polls making up the RCP average. Nine polls make up the average as of this writing, and Biden has five of them. Trump has three, and one is tied.
Issues & Insights poll results surprise me … they are an openly far-right news source. For example, there’s a donation blurb on the right side of their pages that reads “Help us fight for honesty in journalism and against the tyranny of the left.” There is also a “Sites We Like” section at right which lists sites such as the Babylon Bee, RedState, The Daily Caller, Zero Hedge … and RealClear Politics 
Biggest takeaway to me, starting about midway through the article, the section that’s titled “Another poll, another media failure”: Biden is not getting wiped out in the swing states as badly as a recent WSJ set of polls has suggested. Bad methodology (again) in polling Latin-American voters exclusively in English, and the curious involvement of Trump’s 2020 campaign pollster Tony Fabrizio.
More detail in the link.
In the link in the immediately prior post I find:
By definition, a cheerleader is someone whose analysis is biased.
But the consensus of the polls was on target. From 538:
The Polls Were Historically Accurate In 2022
Your source complains about pollsters who are not multi-lingual supposedly undercounting Biden support. I don’t know the technical issues there. But I have a strong suspicion this didn’t change since 2020, when the Republican ticket beat the polls, or 2022, when polls did quite well.
Your link is a classic example of playing the ump. Except – the professional pollsters are not judges whose calling of balls and strikes helps determine who wins the baseball game. No one who reads the New York Times is waiting to see who their pollster thinks will win and then deciding their vote on that basis. So I don’t see what the point is of the cheerleading.
I think it’s to combat the potential (but probably dubious) bandwagon effect.
There is also a similar “doom loop” issue where a candidate is seen as failing and then becomes labelled as a failure. By undercutting the “candidate is losing” argument, pundits and boosters can try to short-circuit this loop.
It’s why one of Biden’s stronger arguments is “I’ve already beat him once and I’ll beat him again”. Pointing out that you defeated your opponent both defies the “Biden is losing” argument and labels the opponent as the loser. Of course, that’s 100% the point of the “stolen election” argument, which a sizable majority of Trump’s supporters believe in.
There probably are people who really believe that there’s truth to be found in majority opinion, even when we know it is a small majority AKA the Bandwagon Effect.
But there also is a real Underdog Effect where you feel smart in going against the crowd, and it feels urgent that we turn this thing around.
I may be more of an Underdog Effect guy.
Regarding this thread, judging the pollsters requires putting both those psychologies aside.
???
Kuo’s mention of the “Democratic cheerleader” is part of a frame story, not part of the analysis. See below for context.
I wasn’t done shaking my head before I read an interview in the New York Times of one of my favorite Democratic cheerleaders, Simon Rosenberg. He had accurately predicted “no red wave” in 2022 based on the data he was seeing and against the consensus of the polls. The interview was mostly fine, but then the Times had to bring up a Wall Street Journal poll out yesterday showing Biden losing to Trump in the swing states. The Times even ran a smug story about how the results “echo other recent surveys, including a series of New York Times/Siena College polls in six battleground states last October.”
“Against the consensus of the [2022] polls” are Kuo’s words. Kuo might or might not be wrong to impugn the 2022 midterm polls.
Rosenberg himself, FWIW, did not see his prediction as opposing the consensus of the midterm polls – he saw it as opposing media coverage of the midterm polls (from Vox.com, 11/27/2022):
In the months leading up to the [2022] midterms, many pundits and politicians thought that Republicans had momentum enough for big gains at the state and federal levels, enough to count as a “red wave.” But veteran Democratic strategist Simon Rosenberg is one of a few voices in Washington who, despite President Joe Biden’s sagging approval ratings and polls that showed Democrats playing defense on inflation, remained optimistic about the party’s prospects and who was ultimately vindicated by a strong performance.
Rosenberg — who has previously advised the Democratic Congressional Campaign Committee and is the president of the progressive think tank NDN — says he’s not in the business of predictions. But he thought that the available data consistently pointed to a competitive election, and he became a self-described “info warrior” on Twitter trying to convince the pundit class of that. He believes that, unlike in 2016 and 2020 when polling failed to register Trump’s strength as a candidate, this time around, it was the media analyzing the polls who got it wrong.
“There was a massive media failure this cycle,” he said. “The failure that just took place is more grave than the polling error [in 2020] because there were a lot of really smart people who basically misled tens of millions of people through their political commentary in the final few weeks.”
Further about the 2022 midterm polling and the predicted “red wave” from the New York Times (from 12/31/2022). The NYT rides the line a bit and allows that “traditional nonpartisan pollsters” were reasonably accurate. But:
… a New York Times review of the forces driving the narrative of a coming red wave, and of that narrative’s impact, found new factors at play.
Traditional nonpartisan pollsters, after years of trial and error and tweaking of their methodologies, produced polls that largely reflected reality. But they also conducted fewer polls than in the past.
That paucity allowed their accurate findings to be overwhelmed by an onrush of partisan polls in key states that more readily suited the needs of the sprawling and voracious political content machine — one sustained by ratings and clicks, and famished for fresh data and compelling narratives.
The skewed red-wave surveys polluted polling averages, which are relied upon by campaigns, donors, voters and the news media. It fed the home-team boosterism of an expanding array of right-wing media outlets — from Steve Bannon’s “War Room” podcast and “The Charlie Kirk Show” to Fox News and its top-rated prime-time lineup. And it spilled over into coverage by mainstream news organizations, including The Times, that amplified the alarms being sounded about potential Democratic doom.
Tying all this into the spring 2024 polling climate: Any reason that the partisan pollsters wouldn’t be doing pretty much the same thing today as they did in 2022?
Anyway, Kuo’s favorite cheerleader, Simon Rosenberg, is mentioned several times throughout this NYT piece. Here are my favorite bits:
Worried that the G.O.P.-inflected polls were wrong and were liable to persuade Democratic grass-roots activists to give up rather than go out and knock on doors, Mr. Rosenberg used his podcast and his Twitter account to tell Democrats that their chances [in the 2022 midterms] were better than they realized.
His bullishness earned him ribbing and ridicule. In an August [2022] article calling him “the most optimistic Dem online,” Politico noted that at times it seemed Mr. Rosenberg was pushing his relentlessly rosy view at “profound reputational risk.”
Watching it all unfold from his offices in Northwest Washington, Tom Bonier, the chief executive of TargetSmart, a Democratic data clearinghouse, worried about the damage being done by overly Republican-leaning polls.
He was certain Democrats were on track to have a surprisingly good year, as liberal and moderate voters alike protested the Supreme Court’s striking down of a federal right to abortion and rejected as too extreme many Republican candidates who continued to deny the results of the 2020 presidential race.
But Mr. Bonier feared “the extent to which perception drives reality,” he said in a postelection interview.
Perceptions of a looming red wave bred stories that sought to explain it, theorizing that voters cared more about crime and the price of gas than, say, abortion.
Mr. Bonier saw a devious logic. He said he suspected that G.O.P.-aligned firms were pumping out polls to help Republicans regain momentum that had shifted toward the Democrats.
“That was the point of this red-wave polling surge,” he said.
As Mr. Bonier and Mr. Rosenberg pleaded with Democrats not to let unreliable polls dissuade them from donating to campaigns or knocking on doors, they repeatedly ran into skepticism from people citing the averages on RealClearPolitics and FiveThirtyEight.
Still, the NYT piece was fair to FiveThirtyEight, not merely heaping dung onto them. The next two paragraphs:
But the averages were being affected by a widening imbalance between a dwindling number of reliable, reputable nonpartisan polls, and a proliferation of questionable surveys.
FiveThirtyEight itself flagged the imbalance: “Compared with past cycles, polls in 2022 are more likely to be sponsored or associated with partisan sources,” an article on the site said in October [2022]. “This is a problem because partisan polls tend to be more inaccurate.”
Rasmussen has never had a poll favoring Biden. Not even just before he won. I ignore Rasmussen, the site does not. It is part of the average.
Never say never.
Rejoice! Nate Silver will be running his Presidential election prediction model again. First version will drop in a month or two. Since he’s doing it himself it will be updated weekly or so rather than every time a poll appears.
The latest Marist poll came in +3 for Biden. It’s the first poll that I have seen this calendar year in which Biden was chosen by more than 50% of the respondents.
Couple that with the Morning Consult’s latest poll – only +1 for Biden, but from over 9,700 respondents – and the RealClear Polling average difference is a barely-there two-tenths of a percent in Trump’s favor 44.6% to 44.4%. Once that Trump +8 Rasmussen poll ages off the list, Biden will likely pull even or be a hair ahead in the average.
The latest Marist poll’s abstract is at least a heartening read, if not a determinative one.
Biden Edges Trump in Two-Way Presidential Contest; Opens Up Lead Among Definite Voters
Despite a still close contest between President Joe Biden and former President Donald Trump, Trump’s support among white voters and independents has splintered. Biden is doing better than he did in 2020 among white voters, and he has eliminated the advantage that Trump had among independents earlier this month. Biden has a three-percentage point edge over Trump among registered voters nationally. In addition to independents, Trump has lost some ground among those who have an unfavorable view of both Biden and Trump. Among those who definitely plan to vote in November, Biden holds a six-percentage point lead. Biden is up five-percentage points in a multi-candidate field.
All caveats about battleground states, etc. still hold.
Also, same link:
“Despite some weak spots for Joe Biden among non-whites and younger voters, he continues to outperform his 2020 numbers among white voters,” says Lee M. Miringoff, Director of the Marist Institute for Public Opinion. “Although it bears watching in the future, right now, a multicandidate field does not benefit Trump.”
aside from election polls, separate Trump and Biden polls find
42% approve
of man who nods and farts at trial How Popular Is Joe Biden? | FiveThirtyEight
39% approve of Biden. Who did not put Israel in its place. Trump would have done it just by telling them.
Plus, Biden is still too old and getting older every day!
Right, if the battleground states follow national trends, as happens often these days, this is also good news for downballot races, which means that the Senate might hold at 50-50, only losing West Virginia, instead of flipping.