Can you trust the polling numbers in the United States in 2020? Pollsters say yes, polling has improved since its failure in 2016 to forecast Donald Trump’s win. And yet, uncertainties remain.
Looking at the political polling data today, it would seem obvious that Democratic presidential candidate Joe Biden is winning the election.
But the memory of Donald Trump’s surprise victory in 2016 still haunts forecasters and pollsters four years later. There’s a level of caution in the wind. Few will venture a prediction.
“If we didn’t have in our memory 2016, we would be near certain that Biden could win,” said Shibley Telhami, a pollster and professor of political science at the University of Maryland.
“Any real objective analysis, including an objective comparison to 2016, would lead you to believe that Trump’s situation is pretty much hopeless,” Telhami told Al Jazeera.
Averages of high-quality surveys – those done by telephone, not online, and with track records of accuracy – show Biden leading in the three Midwest states of Wisconsin, Michigan and Pennsylvania that gave Donald Trump the presidency in 2016.
If everything else remained the same and Biden won those states, he would win the White House. But nothing is ever that simple in politics and Trump is competing aggressively in the Midwest and elsewhere, claiming polls show him leading.
“We’re at 52 percent, they don’t want to report it,” Trump said at a campaign rally in Pennsylvania on October 26.
“You know any poll that comes out positive, they don’t put it on. I’ve been told by pollsters. ‘Why don’t you get it out? It’s a great poll. Sir, we do get it out. They will not report it’,” Trump said, purporting to recount conversations with his campaign advisers.
Trump blames the media for failing to report polls showing him ahead. “These people are the most dishonest,” Trump said in Pennsylvania, pointing to media cameras covering his rally and claiming he’s leading in Nevada and “leading big” in Texas and Ohio, and up four points in Florida.
Obama is drawing VERY small (tiny) numbers of people. Biden is drawing almost no one. We are drawing tens of thousands of people. You’ll see that again today. The Great Red Wave is coming!!!
— Donald J. Trump (@realDonaldTrump) October 27, 2020
Actually, the public polling shows something else. Biden is leading in Nevada and other battleground states – and is effectively tied with Trump in Florida and Ohio while Trump’s lead in Texas has narrowed so much that independent analysts are rating it a toss-up.
“We know that Donald Trump is trying to bring a greater turnout among his core,” said Lee Miringoff, director of the Institute for Public Opinion at Marist College, which is polling voters in 2020.
“Put the tag on them – they are men, they don’t have a college education, they live in the Rust Belt and they didn’t vote last time,” Miringoff told Al Jazeera.
“It’s a gamble because if the turnout is high in other groups – suburban women – even if Trump is successful, he still fails,” said Miringoff, who is comfortable that the polling in 2020 is accurate.
There are two big uncertainties pollsters are grappling with in 2020: the coronavirus pandemic and the massive wave of early voting, which have pollsters guessing at what the electorate will look like on Election Day.
Most of the early voting has tended to favour Biden. And polls show most of the Election Day voting is likely to favour Trump.
“It is a dynamic unlike any we have ever seen,” said Don Levy, director of the Siena College Research Institute, which is producing polls for The New York Times.
Meanwhile, Trump has been drawing large, enthusiastic crowds of several thousand supporters at outdoor airport rallies he is holding in key battleground states, saying it presages a “red wave” of Republican support on Election Day.
“Polling is by definition very uncertain because it is an estimate of an uncertain population, that is, people who turn out to vote,” said Jim Henson, director of the Texas Politics Project at the University of Texas.
“I don’t know, and nobody knows what all the characteristics of the voters will be when all the voting is done,” Henson told Al Jazeera.
The truth is that polling in 2016 was mostly accurate, a post-election review by the American Association for Public Opinion Research found. Hillary Clinton did win the popular vote narrowly, as polls had predicted. But she lost where it matters – in the Electoral College.
Where polling failed in 2016 was in state-level polls underestimating Trump’s support in the upper Midwest among white men without college degrees – now a demographic that pollsters are tracking very closely.
Another overlooked factor in 2016 was that most state polls were conducted a week or more before the election and did not pick up the late-breaking swing in undecided voters to Trump. This time around, pollsters see far fewer voters who have yet to make up their minds about Biden or Trump.
“There have been far more polls by people who are good at what they do in Michigan, Wisconsin and Pennsylvania – the most important three states,” Levy told Al Jazeera.
“If everything stays as it is now, and these polls over the next days show the same and Trump ends up winning Michigan, Wisconsin and Pennsylvania, we’ve got a lot of ‘splaining to do, because right now the averages in those three states are pretty clear,” Levy said.