(CNN)More than half a year after the 2020 presidential election, research is starting to provide a more precise picture of the demographic composition of the 2020 electorate, how various groups voted, and what changed from previous elections. The latest entry, released Wednesday by the Pew Research Center, credits Joe Biden's victory over Donald Trump to shifts in key groups, including suburban voters and political independents, that "on balance...helped Biden a little more than Trump."
The report is based on a survey conducted after November's presidential election. In the months since, researchers went back to validate that participants actually voted by matching them against commercial voter files that aggregate official state turnout records. The results were also weighted to match the final election outcome. It joins an increasing body of election post-mortems, including a report from the Democratic data firm Catalist, as well as data from the 2020 exit polls and VoteCast.
Election polls, as 2020 made clear, aren't precision instruments. The design of these projects helps to mitigate one source of pre-election polling error -- the uncertainty over who'll eventually turn out to vote. They show a general consensus on some findings, such as the existence of stark divides along educational lines. But they diverge in others, such as their estimates of the precise racial composition of the electorate, a reminder that there's still significant uncertainty as there is no single source of definitive truth about who voted and for which candidates in US elections.
The data can also tell a different story based on whether it's used to compare differences between multiple demographic groups in the most recent election, or used to track the way a single group's voting behavior has shifted over time. The new report finds, for instance, that Trump did better with Hispanic voters in 2020 than he did four years prior. At the same time, it's also true that Biden won broad majorities of Black, Hispanic and Asian voters in 2020, while losing the White vote to Trump.
Here's a summary of Pew's findings about last year's election:
GENDER: Female voters chose Biden over Trump by an 11-point margin, according to Pew's data, while male voters were closely divided, giving Trump only a 2-point edge. That's a significant shift from its view of the 2016 electorate, in which Pew found a much bigger gender gap, with women voting for Hillary Clinton by a 15-point margin, and Trump taking an 11-point lead among men. Other sources show a less dramatic change: Exit poll data, for instance, gave Trump an 8-point advantage with men in 2020, only a 3-point dip from 2016. In Catalist's reporting, Biden won 47% of the two-way vote among men in 2020, only a slight uptick from 44% in 2016.
RACE: White voters went for Trump by a 12-point margin, Pew found, while Biden won Hispanic voters by 21 points, Asian voters by 44 points and Black voters by 84 points. Still,Trump's numbers with Hispanic voters were a significant improvement from 2016, when he lost that bloc to Clinton by 38 points. Biden's lead was much stronger among Hispanic voters with college degrees than it was among Hispanic voters who hadn't graduated from college, the analysis finds: he won by 39 points among the former, compared with 14 points among the latter.
RELIGION: Overall, Protestants chose Trump by a 19-point margin, according to Pew, but underlying that figure is a yawning racial divide: White evangelical protestants chose Trump over Biden by a 69-point margin, an uptick from 2016, while Black Protestants went for Biden by an 82-point margin. Catholics were close to evenly split between the candidates, while followers of other religions favored Biden by 32 points. Those unaffiliated with a religious group backed him by a 45-point margin.
AGE: Voters younger than 30 favored Biden by a 24-point margin, Pew found, with those ages 30 to 49 favoring him by a more modest 12 points. Trump had an edge of 6 points among those ages 50 to 64, and 4 points among those ages 65 and older.
GEOGRAPHY: Biden won by a 33-percentage point margin among urban voters and by an 11-point margin among suburban voters, the analysis found, while Trump won by 32 points among rural voters. Biden performed more strongly with suburban voters than Clinton did in 2016, while Trump improved his performance among rural voters.
EDUCATION: Levels of education, which emerged as an important division in the 2016 election, remained so in 2020, particularly among White voters. In Pew's data, Biden led Trump by 15 points among White voters with a college degree, but trailed Trump by 32 points among White voters who didn't attend college (a slight improvement for the Democrats from 2016, when Trump beat Clinton by 36 points among the same group.) The different sources for election data, however, diverge significantly in their estimates of the educational and racial divides. Exit polls, Catalist and VoteCast data all found Biden winning White college voters by a narrower margin than Pew's analysis does. Other data also showed less movement from 2016 among the White non-college vote.
PARTISANSHIP AND VOTING HISTORY: Both candidates won the support of the vast majority of their party in 2020, Pew found. Biden won Democratic and Democratic-leaning voters by a 90-point margin, while Trump won Republicans and Republican leaders by an 86-point margin. The remainder of the electorate -- independents who didn't lean toward either party, or supported another party -- went for Biden by a 9-point margin, up from a near-even split between Trump and Clinton in 2016. Unlike 2016, last year's election didn't feature unusually high support for third-party candidates. Third-party 2016 voters who voted last year opted for Biden over Trump by a 17-point margin, with only a tenth again voting third-party. Both Biden and Trump brought in new and infrequent voters. According to Pew's data, one-quarter of 2020 voters hadn't voted in the previous presidential election, and 19% hadn't voted in either 2016 or 2018. Those who hadn't voted in either 2016 or 2018 were about equally split between the two candidates; Biden led among those who'd voted in the midterms but not 2016, with Trump ahead among those who'd voted in 2016 but skipped the midterms.
Voters' choice of candidates is only one factor in determining an election outcome -- the composition of the electorate also matters. Pew's analysis found that "despite notable changes in the demographics of the two candidates' coalitions," as well as high turnout and a record level of absentee voting, "the demographic composition of the electorate as a whole in 2020 did not differ much from that of 2016."
What did change dramatically was the way people voted. Amid the Covid-19 pandemic, a record 46% of voters said they'd voted absentee or by mail. Of that group, 43% said it was their first time ever voting that way.
The Pew Research Center surveyed 11,818 US adults in November 2020, 10,640 adults in November 2018 and 4,183 adults in November and December 2016, using the nationally representative American Trends Panel. More details on the survey methodology are available here.