Polls overestimated Trump in the primary. Don’t expect that to translate to November.

Former President Donald Trump didn’t quite live up to his polling numbers in the Republican primaries. Yes, he swept past his GOP opponents, but usually by smaller margins than the polls predicted.

For some Democrats, this looks like a hopeful sign for the general election. Polls show Trump narrowly leading President Joe Biden, too. Maybe, these Democrats reason, Trump is also being overestimated in these surveys.

They shouldn’t count on it.

It’s true that surveys overestimated Trump’s actual vote share in eight of the 10 states where there was enough polling for FiveThirtyEight to produce an average. And after a spate of national polls showed Biden behind Trump, the president’s campaign asserted that polling “consistently overestimates Donald Trump while underestimating President Biden.” Aides cited the results of the presidential primaries and special elections held since 2022 as proof.

But both are misleading indicators. That’s because the reason polling overestimated Republican margins in the primaries and special elections is unlikely to be duplicated in the general election: It was hard for pollsters to know which voters were going to show up.

Using polling to estimate election results requires making informed choices about the composition of the electorate. But the Republican primary electorate is much smaller and harder to predict than the much broader segment of voters likely to turn out in November for the general election.

The former was generally populated by high-information voters — and, in states with open primaries, independents and Democrats crossing over in greater numbers to vote for former U.N. Ambassador Nikki Haley. The latter will be different, much larger and in some ways more predictable, since most voters participate in general elections.

The differences between the two are “apples and oranges,” said Monmouth University pollster Patrick Murray, who attributed Trump’s underperformance to a surge in Democrats and independents participating in the Republican primary. “What we believe happened is they came out in higher numbers than we’d ever seen crossover voters come out before,” Murray said.

“It’s not like we missed it because we’re missing voters who had turned on Trump,” he added, pointing to surveys Monmouth conducted for The Washington Post in New Hampshire and South Carolina ahead of those states’ primaries. “We would still have had them in our general election poll.”

Those general election surveys have been a struggle for pollsters in the Trump era, underestimating his support in both 2016 and 2020. Haley’s better-than-expected showing in the primaries doesn’t mean they’ve solved that problem for the general election — let alone overcorrected for it. It just means more voters who would be open to or even inclined to vote Democratic in the general are participating in GOP contests in the primary.

Consider the biggest polling misses from this year’s primaries.

In Michigan, Trump’s margin over Haley was 15.3 points smaller in the actual results than in the final polling average. Pollsters also had double-digit whiffs in Massachusetts (a 14.3-point difference), Tennessee (11.3 points) and Virginia (20.8 points).

Of these four states, only Massachusetts has party registration, but it also allows independents to vote in the presidential primary of their choosing. In the other three, any voter could participate in the Republican presidential primary. And there’s evidence of some crossover voting. Just six-in-10 Virginia GOP primary voters identified themselves as Republicans, according to network exit polls.

Other prominent misses were in states that had only a small sample of GOP primary polling. In Vermont, there was just one poll: A University of New Hampshire survey that showed Trump 30 points ahead of Haley. She ended up winning the state. Vermont has no partisan voter registration, so Democratic-leaning voters could cast Republican primary ballots to oppose Trump.

But trying to extend the consistent overestimation of Trump in the primaries to the general election — where the latest RealClearPolitics national polling average shows Trump ahead by just under 2 percentage points — is problematic.

The electorates are simply not comparable. Fewer than 1.1 million Virginians voted in Super Tuesday’s presidential primaries, just over 17 percent of all registered voters. Compare that to the 75 percent turnout in the 2020 presidential election, when 4.5 million Virginians voted.

Of course a pollster trying to discern which 17 percent slice of Virginia’s electorate will show up for mostly uncompetitive presidential primaries — and which party’s ballot they will pull — would face greater challenges than polling a general election in which most voters participate.

“For people to over-read or over-interpret the Republican primary results — you’re making a stupid mistake,” said Republican pollster Neil Newhouse. “The Republican primary drew lots of voters who are not Republican primary voters. There is very little you can read into this data that has any impact on the general election at all.”

The Biden vs. Trump polls could still be wrong for other reasons — or at least not predictive of an election that’s still less than eight months away. Trump could be stuck in court in a criminal trial as early as next month. And Biden’s campaign is kicking into gear now that both men have vanquished their opponents.

Plus, maybe it’s not about Trump and the presidential primary: Biden’s campaign, in its statement, noted that the party has also done better than expected in special election results since the 2022 midterms, including Democratic Rep. Tom Suozzi’s more-decisive-than-expected victory in New York last month.

But an analysis from The New York Times suggests a similar shortcoming in trying to connect the performance of polls from special elections to the presidential general election.

That analysis shows voters who said they supported Biden in the 2020 presidential election turned out for special elections at a significantly higher rate than those who said they voted for Trump. The Times’ Nate Cohn wrote last month that the differential turnout rate “explains the entirety of the Democratic performance overall.”

In other words, the special election results only would suggest that the general election polls are wrong if Democrats have the same turnout advantage in November. That’s an unlikely proposition.

None of this means the polls are correctly measuring the electorate in this snapshot in time, either. But if they’re wrong, the fact that Trump didn’t win the Republican primaries by the margins the polls suggested isn’t evidence of that.