By any conventional measure — election polls, approval ratings, economic trends, pandemic death rates — President Trump should be headed for a sound beating in November.

Point this out to any nervous liberal, though, and you’ll probably hear this rejoinder: “But 2016!” Pervasive thinking on the left and right holds that Trump surged four years ago on a wave of invisible support that the pundits completely missed, so of course we’re prone to miss it again.

It‘s a fundamental misreading of what happened in 2016. And the problem with misreading elections is that it makes you more likely to repeat your mistakes.

There’s something validating, on both sides, about the popular narrative of the 2016 election, this idea that Trump has some kind of magnetic, undetected hold on wide swaths of the American electorate.

Liberal activists have long suspected that a huge part of white, rural America is irredeemably racist and misogynist — that Barack Obama’s two convincing election wins were an aberration and social progress mostly an illusion.

And Trump supporters love to promote the idea that his following is vast and ascendant. They’d have you believe Trump ignited some kind of powerful ideological movement in the nation’s overlooked white communities.

In fact, exit polls from 2016 tell a very different story.

Nationally, a clear majority of voters four years ago soundly rejected Trump’s anti-immigrant agenda and anti-inclusive message. Only 13 percent said immigration was the most pressing issue in the country. Only 41 percent bought into his idea of a wall on the U.S.-Mexico border.

Fully 70 percent said Trump’s treatment of women bothered them “some” or “a lot.” More than 6 in 10 said Trump was unqualified to be president and lacked the right temperament. Clear majorities thought Hillary Clinton met both tests.

So what happened? Well, here’s where it gets interesting.

Of the 47 percent of voters who said Clinton wasn’t qualified to serve, only 5 percent voted for her. That’s about what you’d expect.

Of the 61 percent of voters who said the same thing about Trump, however, 17 percent ignored their judgment and voted for him anyway.

In Florida, where Trump won by just over a single percentage point, more than half the electorate found him unqualified, and yet 16 percent of those voters cast a ballot for him. In Pennsylvania, where he won by an even slimmer margin, a stunning 21 percent of voters who said Trump didn’t have the right temperament for the job voted to send him to the White House.

In other words, on the night that marked his apex in political life, Trump’s margin of victory came from reluctant voters who almost certainly thought they were voting for the losing candidate, and who felt confident he’d make a terrible president.

There was never anything like a groundswell for Trumpism. In fact, the election had strikingly little to do with him at all. It was mostly about the intense emotions triggered by his opponent.

In the end, some critical slice of voters who thought Clinton eminently more qualified for the job couldn’t bring themselves to vote for her. And they decided their only option was to take a flyer on a guy who seemed manifestly unfit for the job — and destined to lose in any event.

There are myriad explanations for the astounding depth of this anti-Clinton sentiment; let’s not re-litigate them here. What’s clear is that a lot of white voters thought Clinton didn’t like them very much — and her describing them as “deplorables” didn’t help.

What’s happened since? Certainly nothing to suggest that Trump has managed to convert his accidental victory into a burgeoning movement.

The independent voters who narrowly went for Trump in 2016 deserted him almost immediately, once they saw their worst suspicions confirmed. Trump’s approval ratings have hovered just over 40 percent for most of his presidency, which is basically like a football team that goes 6-and-10 every year.

In the only national referendum on Trumpism since 2016 — the midterm cycle two years later — the president’s party was resoundingly rejected.

Trump is not Ronald Reagan, standing on the shoulders of a resurgent, mainstream conservatism. He’s a fringe figure preaching to a loud minority that can’t fill an arena in Oklahoma, a state he won by more than 35 points in 2016. He emboldens an ugly strain of American extremism, but there’s no evidence to suggest he has swollen those ranks.

According to everything we know about politics, Trump should lose in November. And he will — unless Democrats, failing to learn the right lessons from 2016, insist on cornering those same disaffected voters into backing him again.

Joe Biden isn’t Hillary Clinton. He’s a more naturally gifted politician, and whatever his weaknesses, phoniness and elitism have never been among them. I thought Biden would have beaten Trump walking away had he been the nominee four years ago.

There’s really only one way for Trump to win this election — which is for Democrats to hand him the all-out culture war he desperately wants.

Biden and his team are too smart to fall into that trap. But the same might not be said for the party’s loudest activists, such as the types who want to abolish police departments and who believe that everyone who voted for Trump is a closet Confederate.

If the fall campaign feels like another condescending, “us vs. them” indictment of the so-called deplorables, then the same people who felt trapped into voting for Trump in 2016 may do it again. And the result, despite all the evidence of his political anemia, could look the same.

To borrow from the philosopher George Santayana: Those who misread elections are often bound to relive them.

Watch the latest Opinions videos:

Read more: