Even before the vulgar debut of the phrase “locker-room talk,” or the front-page accounts of sexual assault on airplanes, or the national debate over what constitutes an improper amount of pageant-winner weight gain, there has been much that’s split American men and women apart this year when it comes to electing a president.

To some degree, this was inevitable. The 2012 election saw a historic 20-point gap between the sexes as it concerned support for President Obama and former Governor Mitt Romney. But what makes this year’s election different—beyond the unimaginably lewd headlines and pervasive sense of doom—is the fact that Trump and Clinton have divided not just men and women, but men and women who are married to each other.

According to Democratic pollster Celinda Lake, in previous elections, “Being married constrained the gender gap. A lot of women agreed with their husbands.”

Today, she said, “Married men and married women have been disagreeing more. Married women report being hammered by their husbands to vote the same way, earlier than in the past.

By way of some background: As far as John and Jane Doe are concerned, 1972 and 1976 were the last truly harmonious years for the prototypical American male and female voter. According to Pew, Richard Nixon and Jimmy Carter were overwhelmingly the presidential choices for both.

But by 1980, men and women were split over the election of Ronald Reagan by 10 points: Gallup’s final pre-election poll showed that men preferred Reagan over his Democratic challenger Jimmy Carter by 15 points, while women chose Reagan by just five.

In 2000, with the election of George W. Bush, the sexes were split by a whopping 15 points: Men supported Dubya by seven percent, while women opted for Gore by eight.

This year, the split between men and women is going to be epic: “We will have a gender gap the size of the Grand Canyon,” said Republican pollster Whit Ayers.

The data bear this out. Earlier this month, Nate Silver of FiveThirtyEight published two dueling maps. One showed the likely election outcome, should only women vote:

And the other showed a very different result, should only men vote:

Digging deeper into this gap between men and women, what becomes evident is that the political bonds that come with marriage have been broken—insofar as they are something that have held men and women together at the voting booth.

Ayers explained: [The election] has caused enormous strains within married couples. … And part of the problem is not just preference, it’s that if you’re not for Trump, you have a hard time understanding how any rational human being could be. And the same is true for Clinton.”

Throughout modern American elections, married couples—unlike single men and single women—have tended to vote in unison. A 1989 study by Laura Stoker and M. Kent Jennings observed that “marriage gives rise to a new and shared set of social and economic circumstances,” as well as opportunities to “learn from and influence” each other.

Married couples have also tended to vote Republican (even after controlling for things like income and race)—whether because the GOP has traditionally been seen as the party of “family” values, or because men tend to vote Republican, and their wives have historically voted (according to a 2006 study from the Institute for Social and Economic Research) on the basis of their husband’s economic interests.  It might also be that married people, by virtue of entering into the (until recently) heterosexual and traditional institution of marriage, tend to be a more conservative bunch.

Either way, it’s been working for the Grand Ole Party: In 2012, Mitt Romney beat President Obama among married voters by 10 points. Romney, further, won married women by four points, as did his Republican predecessor John McCain, by six points. The last Democrat to win married women was Bill Clinton in 1996—by just one point.

But because of the incredibly pitched and personal nature of the 2016 race, and in particular the numerous marital controversies surrounding both nominees, large numbers of American spouses are not necessarily on the same page this year.

“In the last few days, I’ve heard really bad insults between husband and wives,” said pollster Frank Luntz, who conducts focus groups for both Fox and CBS. “One of my goals in next few days is to actually do a session with a dozen couples who disagree—I just need to make sure I have a divorce lawyer nearby.”

Counter to the last 20 years of American politics, married women now are looking toward the Democrat in the race. An NBC News poll from early October—four days before the Access Hollywood tape was released—showed married women breaking 48 percent for Clinton, 40 percent for Trump. Among married men, however, Trump beats Clinton 52 to 35. That puts a chasm of 25 points between married men and married women. (To be clear, this spousal split may be a largely white phenomenon––voters of color overwhelmingly support Clinton).

As the hurricane of sexual allegations against Trump has continued to gather strength, Clinton is positioned to further exploit this gap. As proof, witness the surrogacy of First Lady Michelle Obama. Lake said Democratic efforts to reach married women aim “to reinforce the idea of women forming their own opinions, to keep women from voting as their husbands do.”

The first lady’s speech last Thursday, perhaps unsurprisingly, was couched in the collective and feminine “we,” therein distancing women from their male counterparts, and focused squarely at women who understand the more oppressive aspects of male dominance:

“And I have to tell you that I listen to all of this and I feel it so personally, and I’m sure that many of you do too, particularly the women. The shameful comments about our bodies. The disrespect of our ambitions and intellect. The belief that you can do anything you want to a woman.”

Luntz explained that his research has shown that “men still define success by their money and careers. Women—even CEOs and opinion movers—still define success by their kids’ health and happiness. [Trump] hasn’t connected to any of that.” The first lady, however, aimed to do just that: frame the argument against Trump as a matter of parenting.

“Consider this: If all of this is painful to us as grown women, what do you think this is doing to our children? What message are our little girls hearing about who they should look like, how they should act? What lessons are they learning about their value as professionals, as human beings, about their dreams and aspirations? And how is this affecting men and boys in this country?”

If Clinton does manage to carry married women by large margins come November, it’s unclear whether the Democratic Party will finally have changed the voting norms of (heterosexual, white) married Americans to its benefit in the medium-to-long term. Luntz, in fact, argued that the split between husbands and wives has as much to do with Trump’s problems with women as it does Clinton’s problems with men.

“Ninety-eight percent of the gender gap stories are about Republicans having trouble with women,” he said. “Why is that? Why isn’t it at least 40-50 percent of the coverage about why Democrats can’t win men?”

(Undoubtedly, this problem—winning over white, married men—is sure to resurface in the 2018 midterms, when the electorate is decidedly more masculine, older, and whiter, and therefore much less favorable to Democrats.)

But then there’s this: Independent of electoral gains and losses, what of the cultural fallout from this fractious American moment?

Luntz suggested that a divorce attorney might be necessary the next time he sat married couples down to talk politics, and that may have been a joke, but does the Trump-Clinton race signal a new dynamic in American marriages? The very fact that married, white women are striking out on their own in the 2016 polls would seem to signal progress and greater equity—but what does that say about the changing compact of the marriage agreement itself? If, as earlier studies have suggested, the institution has been primarily a union of shared social and economic interests, how might that change if spouses choose radically different routes to advance those interests? Indeed, such spousal diversions would suggest a fundamental disagreement on what those interests even are. (Follow this line of reasoning far enough, and you can begin to understand why Luntz wants that divorce attorney present to discuss this stuff.)

The 2016 presidential race has raised questions about many American institutions—whether media, finance, or government. Marriage, in this decade, has already undergone a radical shift in terms of its basic parameters: Merely to refer to it as a union between a man and a woman is now outdated. And yet, after November, it’s possible that the primary assumptions undergirding the nuptial agreement itself may be in question: The 2016 election may change our expectations about what Americans want out of their marriages—or not. Perhaps it only makes sense that in this season of rancor and discontent, the institution charged with securing partnership between disparate factions may itself be challenged by the vicissitudes of the moment.