Nate Cohn

Sen. Elizabeth Warren’s seemingly inexorable rise and former Vice President Joe Biden’s dogged resilience have been the two major stories of the Democratic primary so far.

Over the next and more intense period of campaigning in the early states, the story could be very different. The two will probably have to confront their vulnerabilities to a greater extent than they have to this point, and they face challengers who are well positioned to exploit them.

Pete Buttigieg’s lead in a recent Monmouth University poll of Iowa is as good a time as any to mark the beginning of a new phase of the race.

Until now, the front-runners in the topline poll numbers have shown few signs of weakness. Biden’s lead has been mostly steady in national polling averages over the last month. Though Warren’s rise in national polls has stalled, it has not seemed to reverse in a significant way.

But their weaknesses have been evident by other measures, including in more specific polling questions, for some time.

Biden has never emerged as a classic establishment front-runner, one with a large number of endorsements and strong fundraising. Now, as the race reaches the point when the establishment front-runner would typically be poised to deploy vast financial resources to fend off attacks and challenges, he finds himself likely to be outspent, even badly, by lesser-known rivals.

Biden may already be suffering the costs in Iowa, where Buttigieg is the clearest beneficiary of his weakness among the party’s elite. Buttigieg, the South Bend mayor, has raised more money than Biden and has steadily gained in Iowa polling, culminating in the Monmouth poll that put him in the lead. (Our New York Times Upshot/Siena poll, conducted earlier, showed Buttigieg rising. He was third, a point behind Bernie Sanders and four points behind Warren.)

The softness of Biden’s elite support may also create a new set of threats, like the new candidacy by Deval Patrick and a potential run by Michael Bloomberg. It’s not clear if either man would win substantial support. But if they do gain traction, it could easily come at Biden’s expense.

Tepid elite support and Buttigieg are not Biden’s only challenges in Iowa. He also has little support from political activists. This has contributed not only to his financial deficit but also threatens his support in a relatively lower-turnout environment like a caucus. Biden was a distant fourth in Upshot/Siena polling, with 13% to 14% of support among voters who were “almost certain” to attend the caucus or those who had attended before.

All of this is separate from another question: whether Biden’s performance on the stump and on the debate stage is strong enough to retain his support as voters tune in more intensively. Biden’s resilience in the polls to this point suggests that these concerns could be overstated, though he does seem to have lost more ground in Iowa, where voters are generally paying more attention than they are nationwide. Of course, it is possible that Buttigieg’s rise, not Biden’s performance onstage, is primarily responsible for the weakening in his support.

Buttigieg’s gain might also come at the expense of another candidate: Warren, who also appeals to the meritocratic-minded, well-educated voters who tend to admire Rhodes scholars or Harvard law professors.

Warren’s rise in the polls has been impressive, but there is no reason to assume it will be durable. In Iowa, 78% of her supporters said they could change their mind, more than with any of her top rivals, according to the Upshot/Siena poll.

She has vulnerabilities of her own, and she has come under greater scrutiny since she began to match Biden in national surveys. Her position on “Medicare for all” funding has given her rivals an avenue to attack, and her standing in national and state polls has seemed to slump over the last month.

Her rivals could also point to questions of electability: Although there is still a long way to go, state and national polls show her underperforming her top rivals in matchups against President Donald Trump.

A substantial share of Warren’s supporters in Iowa (60%) said that a woman would have a harder time defeating Trump and strongly agreed (41%) that sexism was a major factor in Trump’s victory over Hillary Clinton. Warren’s own supporters were likelier to hold these views than supporters of Biden, Sanders or Buttigieg, in Upshot/Siena polling.

And Ms. Warren has another rival for support on her left flank: Sanders.

He has been holding steadily at third in national polls. After a health scare in October, he came back with a strong debate performance and a high-profile endorsement from Alexandria Ocasio-Cortez. His standing in the polls has ticked up since, and he remains highly competitive in Iowa and particularly in New Hampshire, where one recent poll gives him the edge.

According to the RealClearPolitics averages, Sanders is essentially even with his rivals in New Hampshire, and just four points out of the lead in Iowa.

His position — just behind the front-runners, without obvious momentum, and with little news media scrutiny — seems to confer tactical advantages heading into the final stretch. Few candidates have incentive to attack him, as his supporters are among the likeliest to say they won’t change their mind.

And Sanders has outraised all of his opponents so far this cycle, potentially giving him the resources to push to the front of the pack in the early states over the final stretch. Even if he falls short, it could sap Warren’s strength.

It’s unclear whether Sanders’ resources, or collateral damage among his rivals, will be enough for him to claim an early state win, or deny one to Warren.

It seems even harder to know what will happen when the campaign rapidly spreads past the early states to Super Tuesday and beyond.

This article originally appeared in

It May Be a Good Idea for Democratic Front-Runners Not to Look Back

(The Upshot)

It is an article of faith for the activist base of both political parties: An ideologically pure candidate can win by turning out the base and mobilizing a new coalition of voters.

For Democrats, nonvoters have long been a source of hope. They are disproportionately young, nonwhite and low-income, demographics that favor the party. In this year’s Democratic primary, turning nonvoters into voters would seem to offer a path to victory for candidates who might seek to win the general election without persuading moderate swing voters coveted by their party’s establishment.

There may be places in the country where this is true. But in the battleground states likeliest to decide the presidency, nonvoters aren’t overwhelmingly favorable to Democrats, according to an analysis of 623 registered voters who live in battleground states and stayed home in both 2016 and 2018.

These nonvoters participated in recent New York Times Upshot/Siena College surveys of the six closest states carried by President Donald Trump.

They lean Democratic, but only by a bit. The president actually leads Elizabeth Warren, 42% to 41%, among active registered voters who did not vote in either 2016 or 2018. Joe Biden leads this group by three points, and Bernie Sanders leads by four.

Those nonvoters who said that, if they voted, they would vote for any of the three leading Democratic candidates seem no more likely to support progressive causes or a progressive nominee than recent voters. Just 13% of nonvoters who would vote for the Democratic nominee call themselves “very liberal.” Nearly as many of these Democratic-leaning nonvoters, 8%, said they’re “very conservative.”

The results suggest that higher turnout, though helpful to Democrats in their quest to defeat the president, would not necessarily assure them of victory in these states. And the president’s ability to stay close among nonvoters suggests he could remain competitive in a high-turnout election and expand his support beyond his 63 million voters in 2016.

The findings undermine the common assumption that Hillary Clinton lost simply because of low Democratic turnout. This month’s Times/Siena polling does suggest she would have benefited from a higher overall turnout; the top Democratic candidates today lead among 2016 nonvoters by between three and eight points. But the relatively narrow Democratic advantage is consistent with a wide body of evidence that the election would have remained deadlocked, even if Democrats had enjoyed a turnout as favorable as the one they had in 2012.

For Democrats, part of the problem is that the demographics of nonvoters no longer work so clearly in their favor. Nonvoters are less likely to have graduated from a four-year college, and the president excels among less educated white voters who are overrepresented among nonvoters in battleground states.

But the Democratic challenge runs beyond demographics. In general, nonvoters are less ideological and less partisan than demographically similar voters, which weakens the expected Democratic demographic edge.

The president, for instance, has a 17% approval rating among self-identified Democrats who didn’t vote in either 2016 or 2018, compared with an 8% approval rating among Democrats who did vote in either of those years. Trump has a 19% disapproval rating among self-identified Republicans who didn’t vote in either 2016 or 2018 and a 9% disapproval rating among those who did.

The same pattern plays out by demographic group. The most striking case involves black voters: Those who turn out vote all but unanimously for Democrats, but the Democratic advantage is not quite so clear among those who haven’t voted in recent elections.

Biden leads Trump among nonvoting black registered voters by only 59% to 19%, but he is ahead, 89-5, among those who voted in both 2016 and 2018. The sample of black nonvoters is small, at just 72 respondents, but the difference is well outside the margin of error. Warren and Sanders underperform among nonvoting black registered voters by a similar level.

What about whites without a degree, the president’s base? Trump has a 25-point lead over Biden in this group if they have voted in one of the last two elections, 59% to 34%. Those who stayed home, however, back Trump by a smaller margin of 53% to 36%. The pattern holds after controlling for age.

Taken together, the figures imply that a large portion of nonvoters sit out because they do not neatly line up with either party. This fact tends to limit the opportunity available to both parties, but overall it tends to diminish the Democratic advantage that would seem to exist based on their demographics alone.

One exception to this general pattern is among Hispanic nonvoters, who tend to be just as Democratic as their voting counterparts, or even more so.

Overall, Biden led among nonvoting Latinos, 64% to 24%, compared with a 63-30 lead among those who had voted in either 2018 or 2016. The sample of nonvoting Hispanic voters is fairly small, but the finding is consistent with data previously analyzed by The Upshot.

As a result, Democrats enjoyed a far larger edge among nonvoters in our poll of Arizona, where 33% of registered nonvoters were Hispanic, than elsewhere in the battlegrounds. Biden led among nonvoters in Arizona, 55% to 37%, thanks to a 74-12 lead among Hispanics in that group. In the rest of the battlegrounds, Biden had only a one-point lead among nonvoters.

The relative Democratic strength among Hispanic voters may also mean that nonvoters nationwide may be more Democratic than nonvoters in the battlegrounds, since nonvoters nationwide are likelier to be Hispanic. It may add further intrigue to speculation about Democratic prospects in Texas, where nonvoters are even more diverse than they are in Arizona.

Of course, Democrats — or Republicans and the president — could benefit from higher turnout if it drew disproportionately from their pool of voters. The Times/Siena data confirms this has happened before.

The group of voters who stayed home in 2012 and turned out for the presidential election in 2016 was fairly favorable to Republicans. Today, they back Trump by three points over Biden.

Voters who stayed home in 2016 but turned out for the midterms in 2018 back Biden by 17 points in these battleground states. Today, they are just as likely to say they are “almost certain” to vote as those who turned out in 2016.

For Democrats in the primary, the question is not simply who leads among nonvoters, but what kind of candidate could excite them enough to lure them to the polls.

The results do not offer particularly clear guidance here. On the issues, nonvoting Democrats look about the same as those who vote. But the group looks relatively likely to include voters who have not been animated by the cultural fights of the last half-decade, which might be something of a surprise given their relative youth.

Nonvoters are the likeliest group of Democratic leaners to oppose an assault weapons ban or to support reducing legal immigration to the United States. They’re likeliest to agree that discrimination against whites is as big a problem as discrimination against minorities, even though the group is only 50% white. They’re also likeliest to agree that political correctness has gone too far.

It’s a combination that informs why this group is somewhat sympathetic to the president, even though they said they would probably vote for the Democratic nominee.

Perhaps this shouldn’t be a surprise: If they did care about the same things that animated “very liberal” voters, they probably would have gone to the polls in these crucial states. Eighty-six percent of very liberal registered voters turned out in either 2016 or 2018 in these states, the highest percentage of any ideological group.

That said, there are some signs that Democratic-leaning nonvoters are somewhat likelier to favor a bold candidate who offers something different. In no small part because of their relative youth, they said they would prefer a candidate who promises fundamental, systematic change to American society over one who promises to return Washington to normal. In contrast, Democratic-leaning voters who turned out in both 2016 and 2018 prefer a return to normal.

Young nonvoting Democrats are slightly more likely than young voting Democrats to support fundamental, systematic change over a return to normal. But in keeping with the broader pattern, they offer more conservative answers on the issues than those who vote.

The gap is particularly stark on cultural questions. Among people under 30 who lean Democratic, nonvoters are far more likely than voters to agree that the women who are running for president “just aren’t that likable.”

The pattern also extends to issues like immigration or health care. Overall, young Democratic nonvoters are far less likely to describe themselves as very liberal.

From all this, you could stitch together a case that Democrats have struggled to appeal to less educated nonvoters, particularly among those who are young and black. These nonvoters do not like the president and they are not conservatives, but they do not appear to have been socialized into the cultural norms of the party’s liberal, college-educated, coastal base. As a result, they have not responded to the president’s conduct with the urgency and outrage of their peers. And some of them may even dissent from their generation’s progressive consensus on cultural issues.

When it comes to the candidates, Sanders shows relative strength among nonvoters: He has a 41% “very favorable” rating in the group, compared with 33% for Biden and 30% for Warren. This is at least in part because of Sanders’ longtime appeal to young voters.

Not only is Sanders’ favorability rating the best of these three candidates, but he is also the only Democrat whose favorability rating is stronger among nonvoting Democratic leaners than among those who have voted before. His outsider status and promise of fundamental change, without much focus on cultural issues, might offer at least one clue for how Democrats might appeal to these nonvoters, though it need not be the only one.

Although Sanders’ favorability rating is strong in both absolute and relative terms, it is not an overwhelming endorsement.

Democrats also have an opportunity to draw on a different group of nonvoters: the millions who aren’t registered to vote at all and who stand on the sidelines of U.S. politics.

Their views are not measured here, but a prior Upshot compilation of 2018 polls from the Pew Research Center and Kaiser Family Foundation found similar patterns. Trump’s approval rating among unregistered black voters, for instance, stood at 19%, and at a surprising 47% among unregistered young voters.

None of this precludes the Democrats from benefiting greatly from new voters as they did in 2018.

But the poll offers little reason to think these voters are poised to save them. Overall, 45% of nonvoting Trump supporters said they are “almost certain” to vote, compared with 46% of nonvoting Democratic leaners.

Nonvoters Are a Source of Hope for Democrats. But Maybe a False Hope.

Today’s America is so deeply polarized that it can be hard to imagine there are people who are really not sure whether they want to vote for President Donald Trump or his Democratic rival.

But these “mythic,” “quasi-talismanic,” “unicorn” swing voters are very real, and there are enough of them to decide the next presidential election.

They are similar in holding ideologically inconsistent views, but they otherwise span all walks of life, based on an analysis of 569 respondents to recent New York Times Upshot/Siena College surveys in the six closest states carried by the president in the 2016 presidential election.

These voters represent 15% of the electorate in the battleground states, and they say there’s a chance they’ll vote for either Trump or the Democrat.

They don’t neatly fit archetypes of swing voters like suburban soccer moms. In fact, men are likelier to be undecided than women. And they are not necessarily the white voters without a college degree, particularly in the Midwest, who decided the last election.

The poll adds a new mix of characters to the quadrennial cast of swing voters, like a somewhat conservative, college-educated suburban man who does not approve of the president’s performance, but strongly opposes a single-payer health system. Or a young man, perhaps even black or Latino, who is not conservative on policy but resents his generation’s stringent cultural norms and appreciates the president’s defiant critique of political correctness.

For now, these persuadable voters in battleground states have a favorable view of former Vice President Joe Biden and Sen. Bernie Sanders, but not of Sen. Elizabeth Warren, our polling shows.

Persuadable voters are so powerful because their votes effectively count twice: A voter who flips from one party to the other not only adds a vote to one side, but also subtracts one from the other side’s tally. A wide array of evidence confirms their decisive role in recent elections.

In the last presidential election, millions of voters flipped from Barack Obama to Trump or from Mitt Romney to Hillary Clinton. In the Midwestern battlegrounds, the flood of white, working-class defections to Trump overwhelmed the smaller stream of white, college-educated voters who defected from the Republicans.

There are places like Howard County, Iowa, population 9,000, where Trump improved over Romney by nearly 1,000 votes, while Republican turnout increased by only 22 votes and turnout among unaffiliated voters increased by 50 votes.

One might assume that places like Howard County will again be at the center of the contest. But there is no outsize mass of white working-class or rural voters who voted for Trump among the undecided in the Times/Siena polls.

Instead, the Times/Siena polling suggests that the electorate remains deeply divided along the lines of the 2016 election, with many groups contributing a sliver of undecided voters to the broader pool.

The size of that persuadable pool depends on how they are defined. Although there is reason to think some voters have more of a partisan lean than they realize, let’s call the 15% who are still thinking of voting for Trump or a Democrat the potentially persuadable.

As a group they are 57% male and 72% white, and 35% have college degrees. Most, 69%, say they usually vote for a mix of both Democratic and Republican candidates. Among those who voted in 2016, 48% say they voted for Trump, 33% for Hillary Clinton, and 19% for Gary Johnson or Jill Stein or no one. Those who voted in the midterm election voted for the Republican congressional candidate by 1 point.

These potentially persuadable voters are divided on major issues like single-payer health care, immigration and taxes. But they are fairly clear about what they would like from a Democrat. They prefer, by 82% to 11%, one who promises to find common ground over one who promises to fight for a progressive agenda; and they prefer a moderate over a liberal, 75% to 19%.

Overall, 40% describe themselves as conservative, compared with 16% who say they’re liberal. Forty percent are moderate.

Trump leads Warren, 49% to 27%, among this broadly defined group of persuadable voters, slightly improving on his margin over Clinton. He holds a narrow 43-37 edge over Biden, a slight improvement for the president over the Republican performance in the midterm election but far from matching his tallies in 2016.

What explains the group’s Republican tilt? Some fairly likely Republican voters could be remaining open-minded to an unspecified Democratic nominee, but quickly returning to the president’s side against a known challenger.

Democrats might prefer to focus on a more stringent way to define persuadable voters. Many of those who said they could go either way selected Trump against all of his Democratic rivals, or selected all three Democrats against Trump. They might be less persuadable than they let on. Excluding those voters leaves 9% of the electorate that showed no consistent partisanship — they can be called the truly persuadable.

These truly persuadable voters supported Democratic congressional candidates in 2018 by 8 points and have less developed views on the presidential race. They support Biden over the president, 38% to 27%, but prefer the president to Warren, 37% to 20%. Sanders is in between, with the president leading him, 34% to 32%. This group voted for Trump by a smaller margin in 2016, 37% to 30%, with the rest casting ballots for minor candidates.

Looking at the full pool of potential persuadables, it can be hard to glean any clear insights. But individual demographic groups present a clearer picture of voters pulled in different directions by their ideology, identity, self-interest or attitudes about the president.

The white college-educated persuadable voters, in either the broad or narrow definition, have something in common: They may not love the president, but they are not sold on progressives.

They oppose single-payer health care, 60% to 37%, and oppose free college, 55% to 41%.

They disapprove of the president, but only 32% disapprove of both his performance and his policies.

Steven Basart, 28, is getting his doctorate in computer science and describes himself as a Democrat. Yet he would consider voting for Trump, depending on the Democratic nominee.

If it were Warren, he’d vote Republican, he said: “I think she’s going too far to the left, which would take our country in a bad direction.”

Basart is not a fan of Trump’s personality, but he says it’s overshadowing some of his accomplishments.

“There are plenty of things not to like about Trump, because he says things that are not nice and potentially racist,” said Basart, who is Latino. “I care somewhat about those things, but I mostly just care about policies, because at the end of the day, that’s what affects people.”

The relatively small number of persuadable women runs against the assumptions of most electoral analysts, who have long assumed that women are likelier to be up for grabs than men. Now, men are likelier to be undecided than women across all major age, race and educational groups. Like the white working-class voters who remain solid for the president, it seems many women, particularly nonwhite and college-educated women, remain anchored to the swing they already made in 2016.

Persuadable men and women generally hold similar views on the issues, including on the president. But they are deeply split over an assault weapons ban, with persuadable women supporting an assault weapons ban by a 26-point margin and persuadable men opposed by 18 points — including 42% of undecided men who say they are strongly opposed.

The undecided white working-class voters often seem as if they would be quite receptive to Democrats based on their views on the issues. They support single-payer health care, for instance.

But they approve of the president’s performance by a comfortable 63-32 margin, and they are about as conservative as Republicans on the cultural issues that divide today’s politics. By a margin of 84% to 9%, they say political correctness has gone too far. They say academics and journalists look down on people like them, and agree that discrimination against whites has become as big a problem as discrimination against minorities.

The persuadable nonwhite voters seem to be an unusual group. They are likeliest to be male — 64% are men — and 39% are younger than 35. They back single-payer by the widest margin, 54% to 40%. Those who voted say they voted for Clinton in 2016 and for Democrats in the 2018 midterms.

By a wide margin of 52-32, they prefer a Democratic nominee who would bring fundamental, systematic change to American society over one who would return politics back to normal in Washington.

Yet Trump’s approval rating is positive among these nonwhite persuadable voters, with 50% saying they approve and 44% saying they disapprove. A majority opposes an assault weapons ban. They want a more moderate Democrat, 69-26, over a liberal, even as they demand fundamental change, and 35% self-identify as conservative.

Young nonwhite men have not traditionally been considered a key swing demographic. As a group, they overwhelmingly back Democrats. They are the sort of voters whom Democrats would ordinarily think of as a turnout target.

A disproportionate number of persuadable voters tend to be low-turnout voters as well: 28% didn’t vote in both 2016 and 2018, compared with 17% of those who say there’s no chance they’ll vote for the other side.

There’s a common view in politics that campaigns can make a choice between turnout, in which candidates play to their base and try to mobilize new voters, and persuasion, in which they reach out to swing voters.

It turns out that many of the low-turnout voters are also the persuadable ones. They don’t have the clear, ideologically consistent views that make them a natural fit for either party, and so they are less likely to vote as well.

Meet the Swing Voters Who Might Decide the 2020 Election

President Donald Trump’s approval ratings are under water in national polls. His position for reelection, on the other hand, might not be quite so bleak.

His advantage in the Electoral College, relative to the national popular vote, may be even larger than it was in 2016, according to an Upshot analysis of election results and polling data.

That persistent edge leaves him closer to reelection than one would think based on national polls, and it might blunt any electoral cost of actions like his recent tweets attacking four minority congresswomen.

For now, the mostly white working-class Rust Belt states, decisive in the 2016 election, remain at the center of the electoral map, based on our estimates. The Democrats have few obviously promising alternative paths to win without these battleground states. The president’s approval ratings remain higher in the Sun Belt battlegrounds than in the Rust Belt, despite Democratic hopes of a breakthrough.

The president’s views on immigration and trade play relatively well in the Northern battlegrounds, including among the pivotal Obama-Trump voters.

There are signs that some of these voters have soured on his presidency, based on recent polling. There is also reason to think that white working-class voters who supported Trump were relatively likely to stay home in last November’s midterm elections.

A strategy rooted in racial polarization could at once energize parts of the president’s base and rebuild support among wavering white working-class voters. Many of these voters backed Trump in the first place in part because of his views on hot-button issues, including on immigration and race.

Alone, the president’s relative advantage in the Electoral College does not necessarily make him a favorite to win. His approval rating is well beneath 50% in states worth more than 270 electoral votes, including in the northern battleground states that decided the 2016 election.

And just because racial polarization could work to the president’s advantage in general doesn’t mean that his particular tactics will prove effective. The president’s campaign rally Wednesday night seemed, for a time, to go too far even for him; on Thursday, he disavowed the “send her back” chants that supporters directed toward a congresswoman who immigrated to the United States as a refugee. (By Friday, he was declining to condemn the chants.)

But Trump’s approval rating has been stable even after seemingly big missteps. And if it improves by a modest amount — not unusual for incumbents with a strong economy — he could have a distinct chance to win reelection while losing the popular vote by more than he did in 2016, when he lost it by 2.1 percentage points.

The president’s relative advantage in the Electoral College could grow even further in a high-turnout election, which could pad Democratic margins nationwide while doing little to help them in the Northern battleground states.

It is even possible that Trump could win while losing the national vote by as much as 5 percentage points.

The state of the Electoral College, 2018

The best available evidence on the president’s standing by state comes from the large 2018 election surveys. Their quality is generally high, and unlike most surveys, they have been adjusted to match actual election results, ironing out many potential biases of preelection polls. Although these surveys are nearly 9 months old, the stability of the president’s overall approval ratings means, for our purposes, that they remain a decent measure of the distribution of his support.

Taken together, the president’s approval rating among midterm voters stood at about 45.5%, excluding the voters who did not express an opinion (for comparability, measures of the president’s approval will exclude voters without an opinion).

By state, the president’s approval rating was beneath 50% in states worth 310 electoral votes: the states carried by Hillary Clinton, along with Michigan, Pennsylvania, Wisconsin, Iowa, Arizona and North Carolina. This is not exactly good news for the president, but not as bad as it typically would be given an approval rating of 45.5%. John McCain, for instance, lost states worth 365 electoral votes in 2008 while winning 45.7% of the vote.

The most important measure of the president’s strength in the Electoral College, relative to the national vote, is the difference between the national vote and the “tipping-point state” — the state most likely to push a candidate over the Electoral College threshold.

Wisconsin was the tipping-point state in 2016, and it seems to hold that distinction now, at least based on the president’s approval rating among 2018 midterm voters.

Overall, the president’s approval rating was 47.1% in Wisconsin, above his 45.5% nationwide. This implies that the president’s advantage in the Electoral College, at least by his approval rating, is fairly similar to what it was in 2016.

A closer look at the underlying evidence suggests there’s reason to think the president’s ratings could be higher than estimated in the state. The estimates are based on four measures of the president’s standing, and there is one outlier: the Votecast survey, which places the president’s net approval rating at minus 13, or 43.6% approval. The other three are in close agreement, placing the president’s rating between 47% and 48%.

There is an additional piece of evidence, unique to Wisconsin, that’s consistent with a stronger position for the president: the Marquette University poll, which gave Trump a minus 5 net approval among likely voters in its final poll before the midterms. Over the longer run, the president has averaged a minus 5 net approval among registered voters (not midterm voters) in Marquette polls since October.

In other words, most measures suggest that the president’s rating is higher than 47.1% in Wisconsin. If you excluded the Votecast data and added the final Marquette poll, the president’s approval rating would rise to 47.6% — or a net 4.2 points higher than his nationwide approval.

It is important to emphasize that it is impossible to nail down the president’s standing in Wisconsin, or any state, with precision. But Wisconsin is the pivotal state in this analysis, and a 1-point difference there could potentially be decisive.

One reason that such a small swing in Wisconsin could be so important is that the Democrats do not have an obviously promising alternative if Wisconsin drifts to the right.

In 2016, Florida was that obviously promising alternative: It voted for Trump by 1.2 percentage points, compared with his 0.8-point victory in Wisconsin.

But all of the measures indicate that Florida has shifted to the right of the nation since 2016, at least among 2018 midterm voters. The president’s approval rating in Florida was essentially even, and by our measure, slightly positive. Republicans narrowly won the Florida fights for Senate and governor, and also the statewide U.S. House vote.

The next tier of Democratic opportunities doesn’t provide an easy backstop to Democratic weakness in Wisconsin either. There’s Arizona, where Democrats had a good midterm cycle, but where the president’s approval rating is plainly stronger than it is nationwide or in Wisconsin. The same is true of Iowa or North Carolina, though the president’s standing in those states is somewhat more uncertain in the absence of an exit poll or a high-profile statewide result.

In the end, these states, particularly Arizona, could prove to be a better opportunity for Democrats than Wisconsin. But at least based on this evidence, it would probably be more a reflection of Democratic weakness in Wisconsin than strength elsewhere.

Milwaukee and Miami-Dade

In both Wisconsin and Florida, the president’s resilience seems grounded in two regions: the Milwaukee area and Miami-Dade County.

The president’s average approval rating in the Milwaukee media market stands at 48% — virtually unchanged from what it was in 2016, in a compilation of Marquette University polls since October. His approval has declined in the rest of the state, according to both the Marquette data and the exit polls, which also showed the president holding firm in the Milwaukee area. A similar pattern has showed up in statewide election results, where Republicans have tended to run strongly in the area.

The president’s approval rating in Miami-Dade may even be better than his standing there in 2016, based on three Times/Siena surveys of two districts there, Florida’s 26th and 27th. These polls were also highly accurate, coming within a point of the election results. On average, the president’s approval rating stood at 45.7% among the likely electorate in the two districts — well above his 40.8% share of the major-party vote there in the 2016 presidential election.

At first glance, these regions might seem to have little in common. But in terms of politics, their idiosyncrasies have played out in similar ways.

Both are regions where the Republicans do better than demographics would lead you to expect. Milwaukee is one of the last Northern metropolitan areas where Republicans still rule the suburbs; Miami-Dade is one of the few places where Republicans win Hispanics, in this case Cuban voters.

Both areas were, or still are, represented by major establishment figures in Republican politics: Scott Walker, Paul Ryan, Jeb Bush and Marco Rubio. Many fought hard against Trump in the primary. These areas were some of Trump’s weakest of the primary season — he won 22% in the primaries in Miami-Dade and in Waukesha, Wisconsin.

Clinton improved over Barack Obama in both areas in 2016. The president’s apparent resilience or recovery in these regions contrasts with what has happened elsewhere in the country. But it is possible that the real anomaly was his weakness in 2016, which was perhaps in part because of the president’s hostility to his prominent skeptics in these areas. The Republican establishment is now unified, if belatedly, behind the president; perhaps these voters have unified behind him as well.

The consequences of higher turnout

Many assume that the huge turnout expected in 2020 will benefit Democrats, but it’s not so straightforward. It could conceivably work to the advantage of either party, and either way, higher turnout could widen the gap between the Electoral College and the popular vote.

That’s because the major Democratic opportunity — to mobilize nonwhite and young voters on the periphery of politics — would disproportionately help Democrats in diverse, often noncompetitive states.

The major Republican opportunity — to mobilize less educated white voters, particularly those who voted in 2016 but sat out 2018 — would disproportionately help them in white, working-class areas overrepresented in the Northern battleground states.

If everyone who was eligible to vote turned up at the polls, the gap between the Sun Belt and Rust Belt would close. Texas, astonishingly, would emerge as the tipping-point state. Wisconsin and Pennsylvania, by contrast, would barely budge.

Of course, a full-turnout election is not going to happen. In recent months, analysts have speculated about a 70% turnout among eligible voters, up from 60% in 2016.

In this kind of high-turnout presidential election, by our estimates, the tipping-point state would drift to the right as people who voted in 2016 but not in 2018 return to the electorate and nudge states like Pennsylvania and Wisconsin toward the president. At the same time, the Sun Belt would drift left. Arizona could overtake Wisconsin as the tipping-point state. But even in this hypothetical high-turnout election, the president’s approval rating in Arizona would be higher than it was in 2018 in Wisconsin. It becomes harder for the Democrats to win the presidency.

In such an election, the tipping-point state could have a net approval rating that is 5 points higher than the president’s national net approval rating, potentially allowing the president to win reelection while losing the popular vote by a wide margin.

2018 isn’t destiny

This analysis mainly covers the opportunities available to both parties; we can’t know which side will take better advantage of them. And it’s important to emphasize that the kind of slight difference in measuring Wisconsin is beyond our ability to discern with great confidence, even using high-quality, calibrated data.

All of this is based on the president’s approval rating — well ahead of the election. Most presidents manage to improve their approval rating between this point and the election, particularly with a strong economy. But unforeseen events could also hurt his approval rating; it is even imaginable that the president could go too far on immigration for some of his more moderate supporters.

If the president’s ratings improve, the crucial question will be where. The answer is likely to be influenced by the contrast he can draw with his still-undetermined opponent.

Democrats could nominate a candidate who tries to win the presidency by mobilizing a new, diverse coalition with relative strength in Sun Belt states, while making little or no effort to secure the support of the white working-class voters with reservations about the president.

The Democrats could certainly win in the Sun Belt states, even in Texas. Perhaps this kind of Democrat could generate such a favorable turnout that it helps the party even in relatively white states.

But it’s also a strategy that would tend to increase the risk of a wide gap between the Electoral College and the national vote. It’s also hard to see how it would be the easier way forward for Democrats, at least as long as the president’s approval rating in the Rust Belt remains so much lower than in the Sun Belt states.

Of course, the campaign season has barely begun. The election could wind up being a simple referendum on the president, and his approval ratings suggest he could lose, perhaps even decisively. But his relative advantage in the Electoral College could ensure his political survival.

This article originally appeared in <a href="https://www.nytimes.com/2019/07/19/upshot/trump-electoral-college-edge-.html" id="2d2af075-cf23-33fd-ae9b-63e2ef4ed587">The New York Times</a>.

Trump's Electoral College Edge Could Grow in 2020, Rewarding Polarizing Campaign

And just because racial polarization could work to the president’s advantage in general doesn’t mean that his particular tactics will prove effective. The president’s campaign rally Wednesday night seemed at least in retrospect to be too far, even for him; he said Thursday that he disavowed the “send her back” chants that supporters directed toward a congresswoman who immigrated to the United States as a refugee.

When Joe Biden first ran for Senate in 1972, 40% of Democrats believed that whites had a right to segregate neighborhoods.

Today, he seeks the presidential nomination of a very different Democratic Party, one that is increasingly progressive, diverse and well educated. Few analysts consider him a clear favorite, even though his consistent early support in the polls would seem to merit it.

If Biden wins, it will be in part because older Democrats still have the votes to decide the party’s nomination in favor of a relatively moderate, establishment-backed candidate, as they have done in the past. He enters with a lead in the polls because of support from those voters.

His candidacy, of course, faces many questions besides whether there’s still a constituency for his moderate, working-class brand. He is also 76 years old, and he has often been considered an undisciplined, even gaffe-prone, campaigner.

In some cases, Biden’s nearly 50-year record in Washington isn’t merely moderate, but out of step with the mores of today’s Democratic Party.

For most of the latter half of the 20th century, Democrats struggled to hold conservative, white working-class voters together in a coalition with African Americans and the party’s liberal elite. Republicans didn’t make it easy: They tried to peel away the conservative white Democratic rank and file, often with success, by appealing to their views on race, crime, welfare, religion and more.

This was not a coalition of ideologically consistent, progressive voters. It can be difficult for today’s policy-driven activists to imagine, but many were voting Democratic mainly because of identity and interest — they saw themselves as union workers, African Americans, white Southerners, the poor. They were strong, lifelong Democrats mainly because of who they were, not because of what they believed.

What they believed was often very different from what many Democrats believe today.

Nearly 40% opposed interracial marriage when Biden first sought a Senate seat, according to the General Social Survey. Nearly 80% opposed legalizing marijuana and believed same-sex sexual relations were wrong. Overall, three-quarters of Democrats were whites without a college degree, and about half of Democrats were moderate or conservative whites without a college degree.

Biden has been criticized recently for having opposed busing in the mid-1970s. At the time, Democratic-leaning voters opposed busing by 79 percent to 21 percent, according to the General Social Survey. Overall, it was about as popular with Democratic voters as President Donald Trump’s tax cuts are today.

White voters without a degree were still a majority of the party’s voters in the early 1990s, when Democrats were seeking to co-opt Republican messages on welfare and tough-on-crimepolicies. By then, the share of Democrats who believed that whites had a right to segregate neighborhoods had dropped to 15 percent, from 40 percent in 1972. But 15 percent is still a significant share, about as large a share as Hispanics represent today in the party. Self-identified moderate and conservative Democrats outnumbered liberals by 2-to-1.

A Pew Research report from 1994offers a striking look at the Democratic coalition at that time. It was an early edition of Pew’s running political typology series, and it included four Democratic-leaning groups under the headline “The ‘Not So’ Left.” Some groups were characterized as “socially intolerant,” or as holding “strongly conservative views on race and social welfare programs.” Even “New Democrats” rejected liberal views on race.

Today, self-identified liberals equal or even outnumber moderates and conservatives within the party. White voters without a degree represent just 30% of Democratic-leaning voters, according to an Upshot analysis of Pew Research surveys from December 2017 to September 2018. Pew’s typology now includes an ideologically consistent, liberal group.

There is no doubt that this is not the Democratic Party of the early 1990s. But it is not the Democratic Party of “woke” social media either. It is somewhere in between.

Biden, for instance, recently faced criticism after accusations that he touched some women in ways that made them feel uncomfortable. It was a crisis for his bid, or at least it seemed that way if you closely followed news reports or were active on social media. But a recent Fox News poll found that just 15% of Democrats said they were “very” or “extremely” concerned about the criticisms, while 51&amp; said they weren’t concerned at all.

Today, the Democratic Party is in transition. Overall, it is roughly divided between its past and its future, and split about evenly between white voters with and without a college degree, between voters older or younger than age 50, and between self-identified liberal and moderate voters.

The fairly even splits make it hard to argue that any particular kind of factional candidate, whether a moderate or a democratic socialist, has an inside path to the nomination.

As much as the Democratic Party has changed, there’s still a group of Democrats who hold fairly conservative views on some issues, like the nearly 30% who oppose marijuana legalization or say same-sex sexual relations are almost always or always wrong. Beyond them, there’s a broader group of voters who hold progressive or liberal views but aren’t exactly progressive activists. A majority of Democrats say political correctness is a problem, for example, according to data from the Hidden Tribes project.

None of these examples suggest that Biden will cruise to the nomination. But they at least complicate the case that he will face a particularly stiff electoral penalty for his conduct.

And Biden brings strengths to the race as well. He enters with substantial good will from having been President Barack Obama’s vice president. He starts with the advantage of being seen by many as the most “electable” Democrat in a year when the party’s voters are desperate to defeat Trump.

Ultimately, Biden seems as if he would benefit in this primary campaign if today’s divided Democratic Party is still a bit more like its past than its future.

This article originally appeared in <a href="https://www.nytimes.com/2019/04/25/upshot/2020-biden-bet-older-moderates.html" id="e60e3e2f-197a-3077-a9c3-83549443ad50">The New York Times</a>.

How Does Biden Fit Into a Changing Democratic Party?

Caucuses have long been the bane of the political establishment and a boost to activist-backed presidential candidates, whether it’s Ted Cruz on the right or Bernie Sanders on the left.

But the establishment is getting a bit of a break for 2020. In this cycle, caucuses are dwindling.

Democratic National Committee rules now encourage states to use a government-run primary, where more people participate, and to increase participation in the caucuses that remain. The package of rules changes, including a big reduction in the power of superdelegates, was approved by Sanders allies last year.

Caucuses tend to be dominated by the most motivated, engaged and informed voters, who also tend to be more ideologically consistent. The opportunity to participate is limited; voting usually occurs at a single time, rather than at any point during a day. The format also can burden voters in ways that deter participation: Caucuses can last for hours and require voters to publicly disclose their vote.

Turning out requires extra knowledge, because caucuses are usually held in a different location than a voter’s regular precinct.

Already, the three largest caucus states — Washington, Minnesota and Colorado — have flipped to primaries. So have Utah, Idaho and Nebraska.

Two more caucus states — Alaska and Hawaii — are using party-run, rather than government-run, primaries. This switch can increase participation and turnout to levels somewhat more like in a traditional primary, depending on how they are administered.

That leaves just six caucus states: Iowa and Nevada — the two highest-turnout caucuses, with Nevada being the only caucus the establishment favorite Hillary Clinton won in 2008 — and Kansas, North Dakota, Wyoming and Maine.

Maine is considering the switch, and North Dakota has switched to a so-called firehouse caucus, which is not obviously different from a government-run primary except that it has limited sites.

Nothing is final, but overall fewer than 5% of pledged delegates are set to be awarded by caucuses in the Democratic primary, down from 14% in 2016. About a third of those delegates will be decided by minor overseas contests in U.S. territories, like Guam and the Virgin Islands.

The remaining caucus states are also adopting rule changes that tend to increase turnout, like absentee or even online voting. It is hard to know how much these will dent the traditional advantages of candidates further to the left, but Clinton was also stronger in higher-turnoutcaucus elections in 2016, winning key early races in Iowa (narrowly) and Nevada.

How much will it matter in 2020? Could it help an establishment-backed politician like Joe Biden if he joins the race?

The overall effect seems as if it will be pretty modest.

That might be something of a surprise. After all, the effect of having a caucus is substantial. In the last two Democratic primaries, Clinton fared about a net 20 points worse in caucuses than primaries, even after controlling for their demographic characteristics.

In 2008, Texas had both a Democratic caucus and a primary, with the caucus awarding one-third of the state’s delegates and the primary awarding the rest. Clinton won the Texas primary by 3.5 points, keeping her in the race, but she lost the Texas caucus by 12.5 points on the same day. As a result, Barack Obama, the insurgent candidate, won the most delegates from the state.

Despite the seemingly small overall effect, you could imagine that a victory in a state in 2020 caused by the switch could lift a campaign for a period, affecting public perceptions and the tenor of news coverage.

Still, caucuses were worth only a net 3 percentage points overall to Sanders (in 2016) and to Obama (2008) in the pledged delegate count (around a 20-point margin on 14% of delegates in each case).

Nate Cohn

Articles written by the author

It May Be a Good Idea for Democratic Front-Runners Not to Look Back

Nonvoters Are a Source of Hope for Democrats. But Maybe a False Hope.

Meet the Swing Voters Who Might Decide the 2020 Election

Trump's Electoral College Edge Could Grow in 2020, Rewarding Polarizing Campaign

Trump's Electoral College Edge Could Grow in 2020, Rewarding Polarizing Campaign

How Does Biden Fit Into a Changing Democratic Party?

Fewer States Will Have Caucuses in 2020. Will It Matter?

The Electoral College's Real Problem: It's Biased Toward the Big Battlegrounds

Why Texas Is Nearing Battleground Status

Huge turnout is expected in 2020, so which party would benefit?

The gerrymandering ruling and the risk of a monopoly on power

Out of debate, yet very much in contention

Politicians draw clear lines on abortion. Their parties are not so unified

How does Biden fit into a changing Democratic Party?

Can polls be predictive this early? Yes, if old rules still apply

Could Howard Schultz help re-elect the president?