The American electorate hasn’t always been this partisan. When Nixon won the presidency in 1972, about 30 percent of his votes came from Democrats; Reagan got near a quarter of his votes from the same source in 1980. This willingness to cross party lines made possible the landslides of 1964, 1972, 1980 and 1984. Throughout the 1980s and 1990s, however, the two major parties became far more ideologically homogeneous. Liberals increasingly identified as Democrats, conservatives overwhelmingly aligned with the GOP, and presidential elections became much closer.

The election of Donald Trump as president and his subsequent takeover of the Republican Party intensified these partisan differences beyond what previously had occurred. From his first day in office, Democrats intensely disliked him and Republicans — especially his base — revered him. It was thus unsurprising that the 2020 election featured a level of partisan voting that exceeded even that of 2016. Likewise, the outcome was once again decided by a handful of votes in the battleground states.

Why, then, did Joe Biden win and Hillary Clinton lose? Our exploration of YouGov/CBS polling provides an answer to that question. We find that in the decisive states (Arizona, Georgia, Michigan, Pennsylvania and Wisconsin), Biden exceeded his national performance among pure independents (the small fraction of voters that eschew all partisan leanings) and substantially bested Clinton’s 2016 showing among that group. He also managed to convert more former Trump voters than Trump did ex-Clinton voters. Ultimately, however, Biden’s improvement on Clinton’s effort was small, reflecting the extraordinary difficulty of winning a contemporary presidential election by a dominant margin in the Electoral College. Moreover, Biden’s win was in large part a repudiation of Trump, which did not carry over to House and Senate races, where Democrats did not meet expectations.

We use two data sources in our analyses: the final pre-election YouGov/CBS survey and a series of October YouGov/CBS polls of various key states. We recognize that pre-election polls underestimated the number of Republican voters in the electorate. However, since we examine relative levels of support among various groups, the inaccuracies in absolute numbers should not impact our conclusions.

The pre-election YouGov/CBS poll confirms our previous observations about party sorting. Republicans overwhelmingly identified as “conservative” or “very conservative” (79 percent in total). Democrats were more heterogenous, but 68 percent still called themselves “liberal” or “very liberal.” Stark partisan divides on the issues highlight the salience of these ideological differences. Table 1 illustrates how survey respondents answered issue questions by party (with those saying they “leaned” Democrat or Republican included as partisans).

These differences are striking, particularly with regard to seemingly non-ideological issues. Over 90 percent of Democrats are concerned about contracting coronavirus, compared to less than half of Republicans. Nearly 90 percent of Democrats worried about foreign interference in the election, while only about 40 percent of Republicans did the same. On more ideologically oriented issues, the disparities are even greater. With regard to healthcare, Black Lives Matter, protests against police misconduct, and the Electoral College, partisans have almost entirely divergent perspectives.

Given this degree of party-ideology alignment and the near-complete issue disagreement across parties, no one would expect 2020 to produce much partisan defection or split-ticket voting. Nevertheless, in a tight election, even a small number of defections — and the voting behavior of independents — can have a decisive impact. This observation is especially true in the battleground states.

For Biden to flip enough Trump-voting states to capture the presidency, he needed to improve upon Clinton’s 2016 margins. In the battleground states, Biden would prevail if he could 1. improve Clinton’s margin among Democratic voters, 2. do better among independents and 3. pick off more Republican votes than Trump did Democratic votes. Biden managed to do all three in 2020. Table 2 shows the two-party vote by partisanship in states he flipped and the country as a whole. Nationally, Biden held more Democrats (about 97 percent) than Trump held Republicans (about 95 percent) and won over 5 percent of the other party’s voters (compared to Trump’s 3 percent of Democratic voters). Among independents, Biden came close to Trump (winning 48 percent).

However, it was in the decisive states that Biden’s true advantage over Clinton became apparent. He did better among Republicans and slightly worse among Democrats than he did nationally; among independents, though, he managed to outperform his national vote share by nearly 3 percent. While that difference may appear small, remember that nearly 24 million votes were cast in the five states Biden flipped, and that his combined margin was only 279,000 (a 3 percent difference translates to 720,000 votes). Biden’s performance among Democrats — about 96 percent nationally and in the flipped states — also substantially improved on Clinton, who received slightly less than 90 percent of Democratic votes. The updated national exit poll supports our analysis — Biden benefited more from partisan loyalty than Trump, won the independents, and captured a higher percentage of Republicans than Trump did Democrats.

Biden’s coalition in the battleground states also included significant numbers of 2016 nonvoters and Trump voters. In the five states he flipped, as well as Florida, Iowa, North Carolina and Ohio, Biden received about 20 percent of his votes from 2016 nonvoters and about 6 percent from 2016 Trump voters. By contrast, Trump received only 13 percent of his vote from the former group and about 3 percent from 2016 Clinton voters. Indeed, our analysis of the battleground state polls shows that Biden won vote-switchers by a decisive 2-to-1 margin. These Trump-to-Biden voters were overwhelmingly concerned about covid-19, with about 82 percent rating it as a “major factor” in their presidential pick. By contrast, Clinton-to-Trump voters were considerably more worried about the economy (about 85 percent rated it a major factor).

Despite the apocalyptic rhetoric surrounding the 2020 election, the results only deviated slightly from the 2016 contest. Voters were somewhat more partisan this past cycle; Biden improved upon Clinton’s performance, especially among independents; and the pandemic soured enough former Trump supporters for Biden to carry critical states. The close presidential result plus the down ballot Republican victories suggest that hyper-partisan elections are not going away soon.

David Brady is the Davies family senior fellow at the Hoover Institution, as well as the Bowen H. and Janice Arthur McCoy professor of political science in the Stanford Graduate School of Business.

Brett Parker is a JD/PhD student and Stanford University and a research assistant at the Hoover Institution at Stanford.

This article is part of a seminar on the 2020 election and its aftereffects organized by the Stavros Niarchos Foundation Agora Institute at Johns Hopkins University. See below for the complete list of contributions to the seminar.