“We can deny some things, but it’s hard to prove for sure what happened,” said Josh Clinton, a professor at Vanderbilt University and chairman of the association’s 2020 election task force. “Based on what we know about voting, what we know about politics, we have some good doubts about what is going on.”
From political campaigns to the media, those “prime suspects” will offer no consolation to opinion pollsters and their dependents. Probably – if far from certain – the culprit for the off-giller voting results is that key groups of people did not respond to the vote in the first place.
Low response rates are a major source of concern For pollsters for more than a decade. But the politicization of the vote in the Trump era — including the former president’s referendum, who falsely claimed that he did not like the referendum results or deliberately suppressed the GOP’s response to voter turnout – may have been counted by some sections of the Republican referendum.
But pollsters say they can’t believe this is the main reason, because you never know who you’re talking to.
This makes it much harder to fix than to find problems with the ballot Four years ago, Which mostly focused on adjusting the polls in view of Trump’s popularity and weakness with college graduates among voters who do not earn college degrees.
“It seems plausible to the working group that the Republicans who run in our election may be different from those who support Republican candidates who do not participate in our polls,” Clinton said. “But how to prove it?”
The first task of the working group is to evaluate the performance of the 2020 general election. In that action, the ballot received a failed standard. Although the national elections have been the worst in four decades, the statewide elections for the Presidential, Senate and Cubernatorial races have been the worst they have ever been (20 years).
According to the report, national ballots for the presidential race held in the final two weeks of the election were averaged 4.5 percent, while state elections were just over 5 points. Most of the errors were in one direction: looking at the vote difference, the national ballot was now very favorable to President Joe Biden by 3.9 points, and the state election was very favorable to Biden by 4.3 points.
Rather than underestimating Biden’s support, most of the errors came from underestimating Trump’s support. Compared to the final election results and the poll numbers for each candidate, Trump’s support was underestimated by an average of 3.3 points, while Biden’s outperformed by one point – a closer, more decisive, race that looks like a solid Biden lead.
This is not just a Trump effect. The polls for the Senate and the governor’s races were even bigger: 6 points on average.
“Within the same state, the voting error in senatorial contests was often larger than in the presidential race,” the AAPOR report says. “Poll margins generally suggest that Democratic candidates will perform better and Republican candidates will perform worse compared to the final certified ballot, regardless of whether they run for president, senator or governor.”
Never put your head and shoulders above others. According to the report, there were only “minor differences” as to whether the phone or the internet was using a mixed system, including texts and smartphone applications – or whether they were randomly contacting voters from the list of registered voters. “Every interview system and every sample system exaggerated the Democratic-Republican difference compared to the final certified vote difference,” the report said.
After the 2016 election, AAPOR’s autopsy blamed that year’s voting errors on a variety of factors. First, the large number of undecided voters measured in the polls rallied proportionally toward Trump at the end of the race, which gave him an immeasurable advantage over before.
But those who decide late on the error of 2020 cannot be blamed: only 4 percent of voters are behind the two main candidates in the state elections held in the final two weeks, and voters who decide late exit polls are equally divided between Biden and Trump.
Another issue of 2016 – which many pollsters failed to weigh in on education – did not go unnoticed last year, the report said. Four years ago, many pollsters adjusted their results to get the right mix of voters based on race and gender. But it lost an important, growing dynamic of the electorate: increasingly, white voters with college degrees favored Democrats, while those without college degrees quickly turned to Republicans. Studies show that voters who do not have college degrees are less likely to vote.
However, by 2020, a majority of state polls made changes to get non-college voters on their ballots. But they were still wrong.
Other 2016-style factors were also ruled out: Voters did not lie to voters about who they would support because of some sort of “shameful Trump” theory (otherwise the errors would not be as big in downhill races). It’s not that a candidate’s supporters did not vote (proof of last year’s race record). Nor can it underestimate the number of voters who voted against what was shown on election day (polls often hit those divisions).
The report is clear that the 2020 vote did not occur. But it says it “seems impossible with available data to determine why the polls exaggerated the Democratic-Republican difference compared to certified votes.”
The most plausible – yet unproven – theory is that the voters who reach the polls are fundamentally different from those who do not. Trump’s rage that the polls are “fake” or worse exacerbates that problem.
“If the most pro-Trump voters are less likely to go to the polls, the voting error could be explained as follows: Self-identified Republicans who choose to respond to the vote are more likely to support Democrats than Republicans,” the report said. “Even if the right percentage of self-identified Republicans voted, Republican differences could create a voting error made and unanswered.”
AAPOR is not the only organization fighting to clear the place where things went wrong. A joint report by five of the largest democratic campaign polling companies released this spring, He said “there is no consensus on a solution” to correct the 2020 errors.
When explanations are elusive, pollsters and their clients find it difficult to make changes to the methods. As fewer Americans are willing to take a 15-minute telephone poll, it is increasingly popular to solicit voter turnout via SMS – or text polls. Online voting continues to grow.
Media-appointed referendums are also changing. A spokesman for the Wall Street Journal confirmed to Politico that NBC News and The Wall Street Journal had suspended the voting federation for more than 30 years late last year. Both news organizations have long worked with the two major polling stations on long-distance telephone surveys.
However, without definitive answers on the causes of the 2020 Miss, pollsters are not sure that it will be possible to get it right in 2022, 2024 or beyond.
“Even after seven months of truth, you think you can know exactly what happened,” Clinton said.
“How sure are we that we can fix this in the future? Well, it’s not clear,” Clinton added. “We’ll have to wait and see what happens – this is not a particularly promising position. But I think that’s the honest answer. ”