With nearly all the ballots now counted in the 2024 election, we can fully evaluate the performance of the national and battleground polls this cycle.
The verdict: They weren’t perfect, but they were more right than wrong — especially given the polling industry’s challenges and recent misses.
Let’s start with the presidential horse race numbers. The final national NBC News poll had Donald Trump and Kamala Harris tied at 49% each, while the national RealClearPolitics average of the two-way contest was Harris 48.7%, Trump 48.6%. And The New York Times’ average was Harris 49%, Trump 48%.
The actual result in the popular vote as it currently stands, which you can expect to change slightly amid the very last counting of ballots: Trump 49.9%, Harris 48.3%.
The 2024 national polls — on average — slightly overstated support for Harris and understated it for Trump. But they did turn in one of the best performances in presidential polling over the past decade, according to historical data from the Pew Research Center.
It was certainly better than polling’s big miss in 2020.
The battleground state polls also weren’t far off the mark, either, although their misses were slightly larger than what we saw in the national polling.
If there’s a perception the polls were wrong in 2024, it’s due less to the actual level of polling error and more to the mistaken impression that political horse race polls can be precise measurements — for example, that 49%-48% or 50%-48% results can perfectly predict who’s going to win a contest and by how much.Given polls’ margins of error, historically low response rates when reaching voters, different assumptions about the electorate and, yes, past polling misses, expecting that kind of precision from political polls has become a fool’s errand.
In other words, all that those 49%-48% and 50%-48% results can tell us is that the race is close.
And that’s exactly what the presidential election proved to be nationally and in the key battlegrounds.
What else the polls got right
Beyond signaling that the 2024 presidential election was competitive and uncertain, the polls nailed the political atmospherics that shaped the contest.
That included an electorate that mostly saw the nation headed in the wrong direction, with an incumbent president, Joe Biden, whose approval rating was stuck in the low 40s — a historical danger zone for the party controlling the White House.
As it turns out, the NBC News Exit Poll found 73% of voters saying they were angry or dissatisfied with the country’s direction, and only 40% approved of Biden’s job performance.
Additionally, the polls foretold many of the key demographic trends that ended up defining the 2024 election, including Trump’s gains with Latino voters.The NBC News/Telemundo/CNBC Latino poll was among the surveys showing those Trump gains well before the election. Many polls also caught on early to Biden and Democrats’ relative struggles with young voters compared to other recent elections, particularly among young men.
Another thing the polls got mostly right in the 2024 election were Democratic downballot candidates who had been consistently overperforming Biden (when he was in the race) and then Harris (once she became the nominee), as well as Republican downballot candidates who had been underperforming Trump in the polls.
That dynamic played out in the election, with Democratic Senate candidates winning in four states — Arizona, Michigan, Nevada and Wisconsin — despite Harris losing in those same states.
What the polls got wrong
But the polls, including the NBC News survey, erred in overstating the size of the gender gap when it came to Harris’ support among female voters and Trump’s backing among men.
According to the exit poll, Harris won female voters by 8 points, and Trump won men by 13 points — a 21-point gender gap that was consistent with recent presidential elections. That was smaller than the 30-point-plus gender gap that the NBC News poll had been showing.
(That said, there was a massive gender gap when combined with race and education, with Harris winning white women with college degrees by 16 points, and with Trump winning white men without college degrees by 40 points — a whopping 56-point gap in the margin between those two different groups.)
And most significantly, the polls once again understated Trump’s support, albeit by a smaller amount than they did in 2020. That came despite polling outfits recalibrating and using new weighting methods in their polls in an attempt not to miss Trump’s voters.
Still, this was the first presidential cycle where many polls showed Trump either ahead or essentially tied — which wasn’t the case in 2016 or 2020, even though that’s how those elections turned out in many contests.
The mixed record on party ID
One of the biggest divides in the 2024 polling were the national polls (including NBC, CNBC and The Wall Street Journal) showing Republicans with an advantage in party identification, versus other national polls that had Democrats with the lead in party ID.
It was the latter group of polls that usually showed Harris ahead of Trump, although within the margin of error.
But in the end, Republicans indeed enjoyed the advantage in party ID, according to the exit poll.
That was a major story in the 2024 election: There were more self-identified Republican voters than Democratic voters.
