Democrats have spent months debating how and why they lost the 2024 election. But the full picture of what happened on Election Day is only now coming into view.

The most authoritative election analyses draw on a variety of different data sources, including large sample polling, precinct-level returns, and voter file data that shows definitively who did and did not vote. And those last figures became available only recently.

The Democratic firm Blue Rose Research recently synthesized such data into a unified account of Kamala Harris’s defeat. Its analysis will command a lot of attention. Few pollsters boast a larger data set than Blue Rose — the company conducted 26 million voter interviews in 2024. And the firm’s leader, David Shor, might be the most influential data scientist in the Democratic Party.

I spoke with Shor about his autopsy of the Harris campaign. We discussed the problems with the popular theory that Democrats lost because of low turnout; why zoomers are more right-wing than millennials; how TikTok makes voters more Republican; where Donald Trump’s administration is most vulnerable; what Democrats can do to win back working-class voters; and whether artificial intelligence is poised to turbo-charge America’s culture wars, among other things. Our conversation has been edited for clarity and concision.

Key takeaways

• Democrats lost the most ground with politically disengaged voters, immigrants, and young people.

• If every registered voter had turned out, Democrats would have lost by more.

• TikTok appears to make its users more Republican.

• Nonwhite moderates and conservatives are voting more like their white counterparts.

• The gender gap among young voters was historically massive in 2024.

• Democrats lost voters’ trust on the economy and cost-of-living.

• Democrats’ most effective message in 2024 was an economically populist one.

• Donald Trump is leaning into the most unpopular parts of his agenda.

• Democratic constituencies are much more vulnerable than Republican ones to AI-induced unemployment.

Before we get into why Democrats lost the 2024 election, let’s talk about how they lost it: Which voting blocs shiftest the furthest right over the past four years?

The most important thing is that we saw incredible polarization on political engagement itself. There’s a bunch of different ways to measure this: There’s how many elections you vote in, or how important politics is to your identity. There’s how closely you follow the news. But across all of these, there’s a consistent story: The most engaged people swung toward Democrats between 2020 and 2024, despite the fact that Democrats did worse overall.

Meanwhile, people who are the least politically engaged swung enormously against Democrats. They’re a group that Biden either narrowly won or narrowly lost four years ago. But this time, they voted for Trump by double digits.

And I think this is just analytically important. People have a lot of complaints about how the mainstream media covered things. But I think it’s important to note that the people who watch the news the most actually became more Democratic. And the problem was basically this large group of people who really don’t follow the news at all becoming more conservative.

If they’re not responding to mainstream media information, where are they getting their views on politics? Is this group reacting to the prices at the grocery store, or firsthand experience of changes in the economy, immigration, or culture?

Most people have to balance their reaction to objective facts in the economy with their preexisting ideological beliefs. A strongly Democratic voter isn’t going to switch to Trump just because they’re upset about high prices. So, it isn’t too surprising that the people with the weakest political loyalties would be the most responsive to changes in economic conditions.

And people who are politically disengaged — like every other subgroup of people this election — overwhelmingly listed the cost of living as the thing they were the most concerned about.

But it can’t just be inflation. Politically disengaged voters went from being a roughly neutral group in 2020 to favoring the Republicans by about 15 points in 2024. But during the Obama era, this was a solidly Democratic group, favoring us by between 10 and 15 points. So there’s also this long-term trend that goes beyond inflation or social media. Our coalition has been transitioning from working-class people to college-educated people.

To move beyond the why, this shift in the partisanship of politically disengaged voters has a really important implication: For most of the last 15 years, we’ve really lived in this world where the mantra was “If everybody votes, we win.” But we’re now at a point where the more people vote, the better Republicans do.

If I understand you correctly, you’re suggesting that Democrats cannot rebuild a national majority merely by juicing higher turnout, since registered voters as a whole were more pro-Trump in 2024 than those who actually showed up at the polls.

Nevertheless, many progressives have attributed Harris’s loss to depressed turnout among Democratic voters specifically. They point to the fact that, between 2020 and 2024, the Democratic presidential nominee’s vote total fell by significantly more than Trump’s tally increased. And they also note that, according to AP Votecast, only 4 percent of Biden 2020 voters backed Trump last year — while a roughly equal percentage of Trump 2020 voters switched to Kamala. So, in their telling, if defections roughly canceled out while a large number of voters went from supporting Biden to staying home, then clearly the problem was inadequate Democratic turnout. So if Harris had focused more on energizing the progressive base, she might have won. What do you think is wrong with that narrative?

Well, the problem with the AP VoteCast data is that it was released the day after the election. There was just a lot of information that they didn’t have at the time. At this point, voter file data has been released for enough states to account for an overwhelming majority of the 2024 vote. And what’s really cool about having that data is that you can really decompose what fraction of the change in vote share was people changing their mind versus changes in who voted.

And when you do that, you see that roughly 30 percent of the change in Democratic vote share from 2020 to 2024 was changes in who voted — changes in turnout. But the other 70 percent was people changing their mind. And that’s in line with the breakdown we’ve seen for most elections in the past 30 years.

The reality is that these things always tend to move in the same direction — parties that lose ground with swing voters tend to simultaneously see worse turnout. And for a simple reason. There were a lot of Democratic voters who were angry at their party last year. And they were mostly moderate and conservative Democrats angry about the cost of living and other issues. And even though they couldn’t bring themselves to vote for a Republican, a lot of them stayed home. But basically, their complaints were very similar to those of Biden voters who flipped to Trump.

The reality is if all registered voters had turned out, then Donald Trump would’ve won the popular vote by 5 points [instead of 1.7 points]. So, I think that a “we need to turn up the temperature and........

© Vox