After impeachment, a coronavirus pandemic and four years of tweets, the early national polls show that the 2020 presidential campaign begins almost exactly where it left off in 2016.
Today, Joe Biden leads Donald Trump by just under six percentage points among registered voters. Nearly four years ago, the final national polls showed that Hillary Clinton led Mr. Trump by around five percentage points among registered voters.
The similarities are even starker on closer examination. The polls depict an electorate that remains split in the same ways, with Mr. Trump and Mr. Biden drawing nearly identical numbers to ones the candidates did four years ago among white voters with or without a college degree.
But not everything is exactly the same, and even modest differences have the potential to alter the race. Mr. Biden shows early strength among older voters; Mr. Trump has made gains among nonwhite voters. There are early signs of an expanded gender gap. All of this could change by November, but all of it could be decisive in a contest with the potential to be closely fought.
Here’s how the coalitions have or haven’t changed in the last four years.
What’s the same
Trump trails by 5 or 6 points
In the compilation of high-quality, methodologically comparable live-interview telephone surveys used for this analysis, Mrs. Clinton led by exactly five points among registered voters in polls conducted after the third presidential debate in October 2016. Mr. Biden, in contrast, leads in methodologically similar polls by 5.8 points.
Of course, state polls struggled in 2016. But the national polls weren’t so bad. The national polls used here showed Mrs. Clinton ahead by 3.7 points among likely voters, or only about 1.6 points more than her final 2.1-point popular vote margin. It’s worth recalling that a 1.6-point shift in Mrs. Clinton’s favor would have probably won her the election, and Mr. Biden will probably be considered a modest if narrow favorite if he enters the election with a five-point lead among registered voters.
A huge split among white voters by college education
The defining feature of the 2016 presidential election was a huge gap between the preferences of white voters with or without a college degree. Four years later, this wide split remains essentially unchanged, with Mr. Trump leading by around 30 points among whites without a college degree and Mr. Biden holding a double-digit lead among whites with a college degree.
It suggests that Mr. Biden, despite his reputed appeal to white working-class voters, has not succeeded in broadly winning back those in that group who flipped from President Obama in 2012 to Mr. Trump in 2016.
It also suggests that Mr. Biden has not made huge gains among college-educated white voters, as some assume based on Democratic gains in suburban areas. Exit polls in 2016, often used for electoral analysis, showed that Mr. Trump won college-educated white voters. But the exit polls were severely biased on this measure. It seems far likelier that Mrs. Clinton won them.
In recent off-year races, Democrats might have won in suburban areas because of a turnout advantage or by persuading traditionally Republican-leaning voters who had already supported Mrs. Clinton to back Democrats down-ballot — not necessarily because they flipped many more college-educated white voters.
A relative Trump advantage in the Electoral College
Mr. Trump’s wide lead among white voters without a college degree means that he has retained his advantage in the relatively white working-class battleground states that have tended to decide recent American presidential elections.
So far, state polls show that Mr. Biden holds only a narrow lead — perhaps around a combined two points — among registered voters in the states likeliest to decide the presidency: Florida, Pennsylvania, Wisconsin and Arizona. It’s consistent with the data from the 2018 midterm elections, when Democrats won the U.S. House vote in Wisconsin and Arizona narrowly — and lost it in Florida — despite winning the national vote by seven points. It’s also consistent with the major 2018 election surveys, which found Mr. Trump’s approval rating about even in the same states.
Mr. Trump’s persistent relative advantage in the Electoral College gives Republicans a chance to win the presidency without the popular vote for the third time in six elections.
What’s different, so far
A Biden surge among the oldest voters
The similarities between the 2016 and 2020 polls are stark, but there are notable differences, too. One is that Mr. Biden holds a wide lead among voters 65 and older, upsetting for now a nearly two-decade pattern of Republican strength among the oldest voters. Over all, Mr. Biden holds nearly a 10-point lead among registered voters in this age group, while Mrs. Clinton trailed among these voters by five points in the final polls.
This is not a small shift. It is consistent across every live-interview poll; it is not easy to explain; and it is fairly new. As recently as the 2018 midterm elections, the president’s approval rating among seniors was above water, and Republicans won the group.
Part of the difference is compositional: Baby boomers now make up nearly 60 percent of registered voters over 65, up from 40 percent in 2016. As a result, Mrs. Clinton was probably fairly competitive among the group of voters who are over age 65 today, even though four years ago she lost those who were over 65. Compositional shifts might explain even more than this, if it turns out that the greater mortality rate of older men compared with women means that Mr. Trump’s older supporters have departed the electorate at a higher rate. But it seems unlikely that compositional change explains it all.
Another factor is a growing bias in polling: Young women respond at a low rate to telephone surveys on cellphones. Pollsters typically adjust their samples to ensure a representative number of women or younger voters over all, but only a few do so in a way that ensures they have the right number of younger or older women. As a result, most pollsters give more weight to women over all, yielding a sample with the right number of women in total but still too few younger ones and, as a consequence, too many older women.
Without any special adjustment, women might represent 60 percent or more of voters over age 65 in a typical high-quality national poll, compared with about 55 percent in reality. This probably expands Mr. Biden’s lead among seniors by about two percentage points — not enough to explain the whole shift but a component.
Together, the polling and compositional explanations might cover nearly half of the shift toward Mr. Biden.
What explains the rest? One possible explanation is Mr. Biden himself. Older voters appear to have a favorable view of him, and it’s not hard to imagine why they might see him in a more favorable light than Barack Obama, Hillary Clinton or even John Kerry.
He is the first white male Democratic nominee since 2004, and he is probably the first one since 2000 or perhaps earlier who won’t easily be dismissed as a liberal. He is not campaigning on the kind of culture war issues, like immigration, racial justice or gay marriage, that have tended to work to conservatives’ advantage with this group for decades. In other words, older voters may like him for many of the same reasons he disappoints young progressives.
Another possible explanation is Mr. Trump — and perhaps especially his coronavirus response. Older people are relatively vulnerable to the coronavirus and relatively insulated from the effects of an economic shutdown. The president’s drive to reopen the economy, and questions about his slow response, may resonate very differently for a retiree who backed Mr. Trump than for a young parent struggling to support children. A recent Morning Consult poll found that Mr. Trump’s approval rating on the coronavirus response was lower among seniors than among any other group.
Whatever the explanation, Mr. Biden’s newfound strength among older voters could be a great asset. They represent an above-average share of the electorate in many top battlegrounds, including Florida. They turn out at high numbers, potentially denting the traditional Republican advantage among likely voters. But it remains to be seen whether Mr. Biden’s strength here will prove durable. Mr. Trump could peel back these losses by refocusing on the cultural issues that have long cemented Republican strength among this group.
Trump gains among nonwhite voters
Mr. Trump, on the other hand, has made gains among nonwhite voters, one of his weaknesses. In recent polls, he holds the support of 22 percent of nonwhite registered voters, higher for him than in any of the late live-interview polls in 2016.
It is hard to break this down further, since many pollsters do not disaggregate nonwhite voters by race, but there are signs that he is overperforming among both black and Hispanic voters and perhaps especially among young nonwhite voters.
The explanation for Mr. Trump’s improvement among this group isn’t necessarily obvious either, but the evidence for it stretches back much further. Nonwhite voters were conspicuously absent from the “resistance” surge in Democratic turnout during Trump-era special and general elections. National polls have long shown that the president’s approval rating was relatively solid, or perhaps even up slightly compared with his 2016 showing, among nonwhite voters, even when his approval rating was far weaker than it is today. .
As with Mr. Biden’s strength among older voters, it’s hard to know whether Mr. Trump’s improvement among nonwhite voters will prove sustainable over the longer run. Either way, it may argue against a surge in nonwhite turnout. Nonwhite voters are usually less likely to vote if they don’t back a Democratic candidate, and even if some wind up returning to Mr. Biden, it seems to suggest they won’t feel any great imperative to turn out and oust the president if they feel comfortable backing him today.
What’s probably different
Biden weakness among younger voters
There are signs that Mr. Biden is running worse among younger voters than Mrs. Clinton did four years ago, but this is not entirely clear.
Pollsters are somewhat inconsistent about how they divvy up voters under 65, making it hard to compile a clear picture of one particular age group. And many age groupings of registered voters are smaller than the 65-and-over group, making the results less precise. All of this is muddled even further by compositional changes, with the Republican-friendly group of Reagan-era Gen-Xers aging into the 50-to-64-year-old category, or the more Democratic-friendly baby boomers aging out of it.
The data is reported most consistently on voters under 35, and over all Mr. Biden appears to fare a few points worse among 18-to-34-year-olds than Mrs. Clinton did. This is a small enough margin that it would be hard to be sure of a change under any circumstances, but the possible biases in the gender composition of young voters in public polls add to the uncertainty. In some polls, men represent 60 percent or more of young voters, even though they represent only 49 percent of young registered voters. This might suppress Mr. Biden’s margin among younger voters by three or four points. It might cover much of the apparent gap with Mrs. Clinton, even as the same theory does not explain Mr. Biden’s strength among older voters in full.
In general, it seems somewhat likely that Mr. Biden is doing worse than Mrs. Clinton among voters under 65, given that he’s doing about the same as her over all while doing much better among those 65 and up. But it’s hard to say with confidence, and even harder to say exactly where Mr. Biden is weak or to what extent. It’s possible that Mr. Biden is faring, say, only two percentage points worse across most age groups under 65; it’s also possible he has a narrower and more pronounced weakness among voters under 29. It’s not clear yet.
A wider gender gap
There was a wide gender gap in the 2016 election. So far, polls show it’s even bigger than four years ago, with Mr. Biden leading by a wider margin among women while Mr. Trump has a wider edge among men.
The shifts are not so large or consistent from poll to poll to know that it’s a sure thing. The shifts are also modest enough that it’s not obvious whether gender or another demographic is driving it. For instance, women make up a larger share among the over-65 group than among younger voters, so Mr. Biden’s strength among older voters may be expanding the gender gap, or it could be that his strength among female voters is helping expand his edge among older voters.
Fewer undecided and minor-party voters
In polls four years ago, a relatively large number of voters said they were undecided or backed minor-party candidates like Jill Stein and Gary Johnson. So far, there’s nothing like it in national polls.
Of course, most national polls aren’t naming third-party candidates. It’s possible that a named Libertarian or Green Party candidate will ultimately win a modest share of support in pre-election polls. But it seems unlikely given the higher favorability ratings of the two major candidates, and it’s even possible that minor-party candidates will struggle to get on the ballot because of the coronavirus. And given the tendency for support among minor-party candidates to dwindle over the course of a campaign, even a modest number of minor-party voters now would not necessarily suggest a similar size vote in November.