This was supposed to be the tightest U.S. election in living memory. So tight, in fact, that we were told it could take days for the winner to be declared. Yet within just hours of polls closing, it was clear not only that Trump had won, but had won big.
The now-president-elect easily passed the required 270 electoral college votes. He also retook control of the Senate, is now on track to win the popular vote, and could also win a majority in the House of Representatives. That’s a fairly clean slate.
On average, Trump reportedly performed around three points higher in the key swing states than polls suggested he would. Plus, he did better with minorities than polls predicted. And remember that Des Moines Register/Mediacom poll days before the election which gave Harris a three-point lead in Iowa? Trump actually won the state with more than 56% of the vote, performing better there than he did in both 2016 and 2020.
The polls also mostly got Trump wrong in 2016. Over in the UK, they were famously off the mark in the 2016 Brexit referendum. And they forecasted a mass of popular support for Keir Starmer’s Labour party in this year’s election, which wound up being the opposite of the case.
Questioned on this repeated gap between polling and reality, YouGov’s chief scientist Douglas Rivers told The Times newspaper in London: “We think the problem is we’re too Democratic,” meaning his organisation failed to weight their polls properly.
We were too Democratic in 2020 and in 2016, and that appears to be consistent.
JLP co-founder James Johnson also said that pollsters continue to underestimate Trump’s support among working-class voters “because they’re busy.”
These are blue-collar workers. They’re not hanging around to answer polls. That’s why we thought that game process or when they’re doing their online shopping was a good way to reach them. These are people with busy, hectic lives who are struggling to make ends meet.
Johnson added that phone surveys—like the laughable Iowa poll—are skewed towards older, liberal women who are “willing to have long chats with you on the phone.”
Shortly after Iowa was called for Trump, American historian Rick Perlstein bashed polling as “a very compromised enterprise,” adding that “it would be great to see people start ignoring it.”
Perlstein noted that even when pollsters get it wrong, as they often do, they insist that each error is an “outlier”; that there was a “huge margin of error.” In other words, “it’s always been heads-I-win-tails-you-lose.”
And it looks as though that won’t be different this time around. New York Times reporter Kaleigh Rogers wrote the day after polls closed that “even before former president Donald J. Trump emerged as the winner in the presidential race, the numbers on election night showed signs of a victory … for polling.” Right.
Rogers later qualifies that “for a morning-after, back-of-the-napkin analysis, it’s looking good for the polls, especially after major misses in 2016 and 2020” (emphasis added), which is hardly saying much.
Pollsters can, of course, decide themselves whether they wish to change their ways after once again proving their uselessness (other than in handing newspapers easily clickable stories). But whether voters continue to read anything into what they say is another question altogether.
