What new research tells us about how Trump won in 2024

President Donald Trump’s 2024 victory over former Vice President Kamala Harris was fueled by “a voter coalition that was more racially and ethnically diverse than in 2020 or 2016,” as well by an advantage among voters who didn’t turn out for the previous election, according to a report released Thursday by the Pew Research Center.

Pew’s analysis, which combines survey data from its in-house panel of poll-takers with information from voter records, contributes to a more clearly emerging picture of the 2024 electorate.

It finds that about three-quarters of eligible voters in the U.S. made the same decision in 2024 that they did in 2020, whether that was voting for the Republican or the Democrat, choosing a third-party candidate or sitting out the election altogether. But one-quarter made a different choice – enough to return Trump to the White House.

Trump held onto 85% of his 2020 voters, the report finds, while Harris retained a smaller 79% of former President Joe Biden’s supporters. Compared to 2020, Trump won a higher share of the vote among Hispanic voters (48%, up from 36%), Asian voters (40%, up from 30%) and Black voters (15%, up from 8%).

“These shifts were largely the result of differences in which voters turned out in the 2020 and 2024 elections,” the authors of the Pew report conclude. “As in the past, a relatively small share of voters switched which party’s candidate they supported.”

Fifteen percent of 2020 Biden supporters and 11% of 2020 Trump supporters didn’t vote four years later, their analysis finds. Trump also won about 5% of Biden’s 2020 supporters, while Harris took about 3% of voters who supported Trump in the previous election.

And while most eligible voters who didn’t cast a vote in 2020 stayed home again last year, those who did decide to vote in 2024 broke for Trump over Harris, 54% to 42%. Adding in people who were too young to vote in the last election, the margin is slightly narrower.

Pew’s analysis is based on the results of a survey conducted just after November’s presidential election. Like all surveys, its results offer an estimate of voter behavior rather than an attempt at pinpoint precision. That’s why different post-election analyses may diverge in some findings about the electorate, even when they converge around a general consensus.

The new analysis, like a report last month from the Democratic-aligned data firm Catalist, incorporates fresh sources of data: information from commercial voters files that aggregate official state turnout records. Pew’s analysis matches that voter file data with responses to their survey – and because its polls are conducted using a panel of respondents who answer multiple surveys over time, researchers there can often track specific individuals’ voting patterns.

Catalist’s report similarly found that voters who turn out irregularly played a key role in Trump’s victory. Since non-presidential elections typically see lower turnout, that could also have potential implications as the parties begin gearing up for the upcoming midterms.

“There’s definitely some evidence that this shift in Democrats doing better among more consistent voters may have some downstream impacts,” said Hannah Hartig, a senior researcher at Pew Research – although she noted that, with a long way still to go until the next election, it’s too early to know how that may play out.

A few more takeaways from the Pew report:

• Trump also improved his numbers among male voterswho split for Trump by a 12-point margin in 2024 after dividing closely between the candidates in 2020. There was especially sharp movement among male voters younger than 50 – while they were about evenly split last year, that marked a swing from a 10-point preference for Biden in 2020. Both Pew and Catalist show Democrats losing more ground among male voters than female voters, while exit polling and post-election data from Votecast found that erosion across gender lines.

• Education remains a major fault line in American politics. College graduates who voted in 2024 broke for Harris by a 16-point margin in Pew’s data, while those without degrees broke for Trump by 14 points – although both those findings represent an improvement for Trump from his 2020 numbers. That education gap persisted among both White and Hispanic voters, while Black voters didn’t divide significantly along educational lines. Catalist’s report found similar educational trends, but charted somewhat less of a divide among Latino voters, while exit polling and VoteCast had showed college graduates’ preferences remaining more stable.

• Naturalized citizens of the U.S. made up about 9% of last year’s electorateaccording to Pew. And in 2024, they were closely divided, with 51% backing Harris and 47% backing Trump. By contrast, in 2020, this group broke heavily for Biden.

The design of Pew’s study also allowed them to check in with nonvoters: adults who were eligible to vote, but weren’t a part of the 64% who actually turned out. In the past, this group typically leaned Democratic: asked whom they would have preferred if they had voted, 2020 nonvoters favored Biden over Trump by an 11-point margin. But in 2024, nonvoters were closely split, with 44% preferring Trump and 40% Harris.

“If somehow something magic had happened and everybody who’s eligible to vote had actually showed up, not only would it not have helped the Democrats and Harris, it might have actually pushed Trump’s margin up slightly,” said Scott Keeter, a senior survey advisor at Pew.

The Pew Research Center surveyed 8,942 US adults in November 2024, using the nationally representative American Trends Panel, including 7,100 voters who were able to be matched against a voter file. Results among the full sample of validated voters have a margin of error of +/- 1.5 percentage points. More details on the survey methodology are available here.

For more CNN news and newsletters create an account at CNN.com