As I try to understand public opinion in yet another presidential election year with former President Donald Trump as the Republican nominee, I see an anomaly.

On one hand, the polls look very much like the 2020 and 2016 election results. Trump trails Vice President Kamala Harris by 1.9 percentage points in the RealClearPolitics average of recent polls, for example. That margin looks a lot like the 2.1-point popular vote margin for former Secretary of State Hillary Clinton in Nov. 2016 and is not qualitatively different from the 4.4-point popular vote margin for President Joe Biden in Nov. 2020.

On the other hand, current poll numbers, representing interviews conducted through the end of August, showed Democrats significantly further ahead in those earlier years — Clinton by 3.9 points, Biden by 7. The differences are even starker in target states such as Pennsylvania, Arizona, Michigan and Wisconsin.

Depending on how you resolve this anomaly, you will be expecting quite a different experience on Nov. 5. If you concentrate on the similarity between today’s polling results and the last two actual election results, you’ll expect yet another excruciatingly close election, with the electoral vote majority probably determined by which candidate carries two or three states by tenths of a percentage point.

If you concentrate on Democrats’ smaller polling margins over Trump this year as compared to 2016 and 2020, however, you’ll expect him to run considerably better this time, perhaps winning the popular vote and probably carrying the electoral vote without difficulty.

So which is the better view? Is he really running better this time, something sufferers from Trump Derangement Syndrome must have a hard time believing, or are we just looking at the same old same old?

There’s evidence for that latter view. I don’t see significant changes in assessments of Trump’s character. He compared favorably on some traits with the flagging Biden, but Democrats’ and dominant media’s month of joy over Harris’ succession has eliminated, at least temporarily, some of the former president’s advantages.

On top of that, there’s little evidence that the close balance between the parties is moving toward Trump’s party. In Senate races, Republican candidates are leading in polls in only two of half a dozen Democratic-held seats that have seemed vulnerable.

In the generic vote for the House of Representatives, Democrats are currently up 1.2 points, with neither party leading by more than that for the last six months. Those numbers point to very narrow majorities for one party or the other in both houses, as in the Senate and House elections in 2020 and 2022.

Finally, the gap between 2016 and 2020 polls and actual election results was greatest, as Republican pollster Patrick Ruffini argues, in closely contested Rust Belt states where polls undercounted whites without college degrees outside million-plus metropolitan areas. Since 2016, pollsters have been trying to cure this defect, and perhaps this year they have. If so, the results this year will be closer to the polls that, at the moment, suggest a close race.

If not, however, and as improbable as it may seem to TDS sufferers, Trump is actually a stronger candidate than he used to be. Polling suggests voters have a better retrospective view of his presidency than of Biden’s — or, as Republicans call it now, the Biden-Harris administration.

Certainly, on two highly visible issues, inflation and immigration, basic metrics make it easy to argue, and hard to deny, the superiority of the Trump record over Biden’s. That has led to Harris’ ludicrous proposals for price controls on groceries, and Democrats’ and sympathetic media’s pathetic attempts to deny she ever had any responsibility for border policy.

It is also widely appreciated, and substantiated by data, that real-dollar earnings and wealth, particularly among those in the lower half of incomes, increased more during Trump’s years than during Biden’s.

Trump supporters have been noting as well that Russia did not advance in Ukraine, and Iran-allied Hamas and Hezbollah did not launch attacks on Israel, during his presidency. Some voters may believe this is just a coincidence, but some may believe his policies or temperament deterred America’s enemies — and would again.

Something else may be happening. People may be recoiling from the events of 2020, when “leftism was on the march,” as poll analyst Nate Silver recalled in May, as “Covid-era restrictions … now viewed as going too far” had “government intervening in everyday life in a way it had never had in most Americans’ experience.”

It was “the year went America went crazy,” as I have written, when the videotaped death of George Floyd led to days of violent riots across the country, and Democrats heeded calls to “defund the police.”

That might have been a favorable atmosphere for Harris if she had not ended her presidential campaign in Dec. 2019, after calling for abolishing Immigration and Customs Enforcement, enacting “some form of (racial) reparations,” and banning fracking and, after 2040, gas-powered cars.

Such stances are less popular today, and Harris campaign staffers have been posting that they’re inoperative now, even as the complacent press remain uncurious about whether and why the candidate changed her mind. Or about why the public has.

If Trump should once again end up outpolling the polls, running ahead of his previous 47% and 46% showings and easily exceeding 270 electoral votes, such an ideological recoil would be the main reason. But maybe this time the polls are right.

Michael Barone is a senior political analyst for the Washington Examiner, resident fellow at the American Enterprise Institute and longtime co-author of The Almanac of American Politics.