Did the presidential polls really get it wrong?

Despite all the “dead heat” and “neck-and-neck” predictions, in the end it wasn’t that close.

Donald Trump secured a clear victory in the U.S. presidential election this month. It wasn’t a landslide, but he won comfortably with 312 electoral votes. The president-elect also won the popular vote by 3.2 million votes, the first time a Republican had done so since President George W. Bush won reelection in 2004.

Predictably, open season was soon declared on pollsters, most of whom had forecast an extremely tight race along razor-thin margins. After major miscalculations in the 2016 and 2020 elections, public patience with the whole polling apparatus was wearing thin.

But is such criticism really fair? By this point, there’s a sense that poll-bashing has almost become a national sport. And Robert Preuhs, a Political Science professor at Metropolitan State University of Denver, said much of the mudslinging fundamentally misunderstands how polling works and what it is trying to achieve.

“The key issue is that we expect precise and definitive answers from polls, which were always meant to be estimates of public opinion,” he said. And besides, he added, many polls this time did get pretty close.

“This year’s statewide poll of polls, from RealClear Politics, was off by less than 2 percentage points in all but one of the swing states, and that’s genuinely useful information for voters,” he said. “If I’d told you the result would be 51% to 49% and we ended up with 49% to 51%, that’s not far off at all.”

Useful information

Still, some conspicuous names did terribly in their predictions. On the eve of the election, polling expert Nate Silver ran 80,000 simulated models that showed Vice President Kamala Harris ahead.

Ann Selzer’s universally lauded Des Moines Register poll showed Harris leading by 3 percentage points in Iowa, which turned out to be off by 16 points. (Trump won the state handily.) Even Allan Lichtman, the academic “Nostradamus” who is famous for correctly forecasting presidential elections, drew a dud.

One of the pollsters’ main headaches, especially when it comes to gaining public understanding, is the margin of error. Because polling is based on a sample of all voters, most pollsters factor in a standard margin of error to their predictions, often in the range of 3% to 4%.

RELATED: Are presidential elections bad for your health?

But recent election results have been so close — the anticipated final margin for the 2024 presidential election, for example, is 1.5% — that polls can be accurate within their own margins and yet still make the wrong call in a close race. That has led some commentators to question whether they are helpful at all.

“It all depends on how you interpret ‘helpful,’” Preuhs said. “We all tend to get bogged down in the minutiae of percentage points. But perhaps the most important feature of polls, which everyone forgets, is that they give us a reasonable expectation of the outcome.”

That single element, he argues, has become all the more important in such a deeply polarized society.

“Many Americans today are politically homogenous in both their physical and digital lives — they simply can’t imagine how anyone could vote for the other party’s candidate,” he said. “And in that context, having a national or statewide poll that explicitly states half the country disagrees with your point of view is an important piece of information in itself.”

Aggregated results

With increased scrutiny and some high-profile misses, the polling industry’s reputation has taken a battering in recent years. One study, covering 1,400 polls from 11 election cycles, was particularly caustic.

It found that polls taken a week before elections were accurate 60% of the time. But the killer detail is that at 10 weeks out, the accuracy rate dropped to 50%, a literal coin flip. That sounds damning. But as Preuhs said, it’s also unfair.

“The study authors looked exclusively at individual polls to get their results,” he said. “And one lesson I’ve drummed into my students for decades now is: Never put too much stock in individual polls.”

RELATED: Colorado voters, here’s how your ballot is counted

Any single poll, by itself, can be prone to a certain amount of error. Instead, Preuhs counseled, it’s always better to focus on polls of polls, which aggregate numerous results.

“These are much more reliable,” he pointed out, “since the errors in each individual poll tend to offset each other in the accumulated average.”

Here to stay

When people become overtly critical of polls, which happens a lot, Preuhs is often tempted to ask: What if they didn’t exist?

In such a world, people living in their own political bubble might never expect a loss and be completely blindsided when the other party coasts to a win. “There could be nothing better to fuel a violent revolt,” he said.

Without polls, he added, voters might be so confident of victory that they neglect to vote and allow the other party to win. Or citizens might not know whether a government policy reflected the general will of the people or just lobbying by an influential group.

“Polls are by no means perfect, and sure, some follow better procedures than others,” Preuhs said. “But overall, they provide useful guidance and information for citizens across the whole social and political spectrum.”

And his final piece of advice? Love or hate polls, we should all get used to them.

“Believe me,” he said. “They aren’t going away any time soon.”