“Can you never do that again?” my son texted me on Monday in our family group chat. I had sent a series of photos of my flight in the tiny Cessna Caravan that had just flown my mortal being 120 miles, from Chicago O’Hare International Airport to West Lafayette, Indiana. The nine-seat aircraft, which runs on a single turboprop engine, was so small that the ground crew had to weigh luggage and passengers in order to distribute their weight evenly in the cabin. It was, in short, the kind of plane that makes it easy to fear for your life. By contrast, I hadn’t been concerned at all—and my son had found no cause to worry—about the American Airlines regional jet that I’d taken on the first leg of my trip, from St. Louis to Chicago.

Just a few days later, an American Airlines regional jet collided with a U.S. Army Black Hawk helicopter in Washington, D.C., killing everyone involved: 60 passengers, four crew members, and three service members. The National Transportation Safety Board has said it will take at least a year to identify a final probable cause of the crash. Until then, one can only guess that the aircrafts and their machinery were not themselves to blame. The New York Times has reported that the relevant air-traffic-control tower may have been understaffed and that the helicopter might have been outside its flight path. As Juliette Kayyem wrote for The Atlantic yesterday, a rise in flight traffic has been increasing the risk of midair collisions for years, especially in busy airspace such as Washington’s.

[Read: The near misses at airports have been telling us something]

Statistically, for now at least, flying is still much safer than driving. According to the International Air Transport Association, on average a person would have to travel by plane every day for more than 100,000 years before experiencing a fatal accident. A host of factors has made flying more reliable, among them more dependable equipment, better pilot training, tighter regulations, stricter maintenance standards, advances in air-traffic control, and improved weather forecasting. But the amorphous, interlocking systems that realize commercial flight are hard to see or understand, even as they keep us safe. For ordinary passengers—people like me and my son—any sense of danger tends to focus on the plane itself, because the plane is right in front of us, and above our heads, and underneath our feet, and lifting us up into the sky. A fear of flying makes little sense, because flying is just physics. One really fears airplanes, the aluminum tubes in which a fragile human body may be trapped while it is brought into flight. A machine like that can crash. A machine like that can kill you.

The Boeing 737 Max’s recent string of mishaps, including two fatal accidents in 2018 and 2019, and, more recently, a lost door during flight, are still fresh in the minds of passengers, and history only reinforces the fear. In 1950, a TWA Lockheed Constellation en route from Mumbai to New York crashed when its engine caught on fire and detached. In 1979, another engine detachment on a DC-10 wide-body jet caused the crash of American Airlines Flight 191. In 1988, an Aloha Airlines Boeing 737 lost an 18-foot-long section of upper fuselage on the way from Hilo to Honolulu. Human error—a contributing factor in most crashes, if not their direct cause—can also stem from equipment failure, as in the case of Pan Am Flight 812 in 1974 and Air France Flight 447 in 2009.

Yet the salience of an airplane’s actual machinery has been fading too. For passengers, the experience of commercial flight may be worse than ever, but the planes themselves now seem more reliable and more accommodating (if only so many passengers weren’t packed into them). Twenty years ago, regional flights would commonly use turboprops to transport passengers between hubs and small-to-medium-size cities. These planes were louder and bumpier. Flying in them felt worse, and it inspired more anxiety for that reason.

Are little turboprops actually more dangerous than jets? A direct comparison is difficult, because the smaller planes are often used for shorter flights, and more flights mean more takeoffs and landings—when most accidents occur. But the numbers are somewhat reassuring overall, at least when it comes to commercial flight. (The numbers for general aviation, which includes recreational planes, skydiving operations, bush flying, and the rest of civilian noncommercial flight, are less reassuring.) The NTSB filed investigations into eight fatal aviation accidents in the United States from 2000 to 2024 that involved commercial aircraft with turboprop engines, and 13 for aircraft with turbo fans (the most common passenger-jet engine).

In any case, modern airport logistics, just like modern jumbo jets, have helped build a sense of safety—or at least hide a source of fear. U.S. passengers used to board and disembark their flights from the tarmac with more regularity. This was true of prop planes and bigger jets alike. The shrill whine of turbines and the sweet smell of aviation fuel made the mechanisms of flight more palpable; it reminded you that you were entering a machine. Nowadays, that reality is hidden. You board comfortable, quiet cabins from the climate-controlled shelter of jet bridges.

All of these changes have tamped down the fear of planes to the point that, for many passengers, it will now resurface only under certain throwback conditions—such as when I found myself bobbing over the Hoosier farms in a plane cabin the size of a taco truck. That sort of white-knuckling is a distraction from the truer, more pervasive risks of air travel in 2025. The systemic lapses and conditions that have produced frequent near misses in aviation, and that may have contributed to this week’s accident, now seem likely to worsen under the Trump administration, which has purged the Aviation Security Advisory Committee, fired the head of the Transportation Security Administration, and blamed DEI for a fatal crash.

[Read: Is there anything Trump won’t blame on DEI?]

The nation’s pervasive weakness in aviation safety is genuinely scary, but it’s shapeless, too. It provokes the sort of fright that you feel in your bones, the sort that makes you entreat a loved one to please never fly in one of those again, okay? And yet, I might well have been safer in the cold cabin of a turboprop 5,000 feet above Indiana than I would have been on an approach to an overcrowded, understaffed airport in a quiet regional jet. The plane still seems like the thing that might kill you. Even now, I suspect it always will.

Perhaps you’ve heard: Young people aren’t dating anymore. News media and social media are awash in commentary about the decline in youth romance. It’s visible in the corporate data, with dating-app engagement taking a hit. And it’s visible in the survey data, where the share of 12th graders who say they’ve dated has fallen from about 85 percent in the 1980s to less than 50 percent in the early 2020s, with the decline particularly steep in the past few years.

Naturally, young people’s habits are catnip to news commentators. But although I consider the story of declining youth romance important, I don’t find it particularly mysterious. In my essay on the anti-social century, I reported that young people have retreated from all manner of physical-world relationships, whether because of smartphones, over-parenting, or a combination of factors. Compared with previous generations of teens, they have fewer friends, spend significantly less time with the friends they do have, attend fewer parties, and spend much more time alone. Romantic relationships theoretically imply a certain physicality; so it’s easy to imagine that the collapse of physical-world socializing for young people would involve the decline of romance.

[From the February 2025 issue: The anti-social century]

Adults have a way of projecting their anxieties and realities onto their children. In the case of romance, the fixation on young people masks a deeper—and, to me, far more mysterious—phenomenon: What is happening to adult relationships?

American adults are significantly less likely to be married or to live with a partner than they used to be. The national marriage rate is hovering near its all-time low, while the share of women under 65 who aren’t living with a partner has grown steadily since the 1980s. The past decade seems to be the only period since at least the 1970s when women under 35 were more likely to live with their parents than with a spouse.

People’s lives are diverse, and so are their wants and desires and circumstances. It’s hard, and perhaps impossible, to identify a tiny number of factors that explain hundreds of millions of people’s decisions to couple up, split apart, or remain single. But according to Lyman Stone, a researcher at the Institute for Family Studies, the most important reason marriage and coupling are declining in the U.S. is actually quite straightforward: Many young men are falling behind economically.

A marriage or romantic partnership can be many things: friendship, love, sex, someone to gossip with, someone to remind you to take out the trash. But, practically speaking, Stone told me, marriage is also insurance. Women have historically relied on men to act as insurance policies—against the threat of violence, the risk of poverty. To some, this might sound like an old-fashioned, even reactionary, description of marriage, but its logic still applies. “Men’s odds of being in a relationship today are still highly correlated with their income,” Stone said. “Women do not typically invest in long-term relationships with men who have nothing to contribute economically.” In the past few decades, young and especially less educated men’s income has stagnated, even as women have charged into the workforce and seen their college-graduation rates soar. For single non-college-educated men, average inflation-adjusted earnings at age 45 have fallen by nearly 25 percent in the past half century, while for the country as a whole, average real earnings have more than doubled. As a result, “a lot of young men today just don’t look like what women have come to think of as ‘marriage material,’” he said.

In January, the Financial Times’ John Burn-Murdoch published an analysis of the “relationship recession” that lent strong support to Stone’s theory. Contrary to the idea that declining fertility in the U.S. is mostly about happily childless DINKs (dual-income, no-kid couples), “the drop in relationship formation is steepest among the poorest,” he observed. I asked Burn-Murdoch to share his analysis of Current Population Survey data so that I could take a closer look. What I found is that, in the past 40 years, coupling has declined more than twice as fast among Americans without a college degree, compared with college graduates. This represents a dramatic historic inversion. In 1980, Americans ages 25 to 34 without a bachelor’s degree were more likely than college graduates to get married; today, it’s flipped, and the education gap in coupling is widening every year. Marriage produces wealth by pooling two people’s income, but, conversely, wealth also produces marriage.

Contraception technology might also play a role. Before cheap birth control became widespread in the 1970s, sexual activity was generally yoked to commitment: It was a cultural norm for a man to marry a girl if he’d gotten her pregnant, and single parenthood was uncommon. But as the (married!) economists George Akerlof and Janet Yellen observed in a famous 1996 paper, contraception helped disentangle sex and marriage. Couples could sleep together without any implicit promise to stay together. Ultimately, Akerlof and Yellen posit, the availability of contraception, which gave women the tools to control the number and the timing of their kids, decimated the tradition of shotgun marriages, and therefore contributed to an increase in children born to low-income single parents.

The theory that the relationship recession is driven by young men falling behind seems to hold up in the U.S. But what about around the world? Rates of coupling are declining throughout Europe, as well. In England and Wales, the marriage rate for people under 30 has declined by more than 50 percent since 1990.

And it’s not just Europe. The gender researcher Alice Evans has shown that coupling is down just about everywhere. In Iran, annual marriages plummeted by 40 percent in 10 years. Some Islamic authorities blame Western values and social media for the shift. They might have a point. When women are exposed to more Western media, Evans argues, their life expectations expand. Fitted with TikTok and Instagram and other windows into Western culture, young women around the world can seek the independence of a career over the codependency (or, worse, the outright loss of freedom) that might come with marriage in their own country. Social media, a woman veterinarian in Tehran told the Financial Times, also glamorizes the single life “by showing how unmarried people lead carefree and successful lives … People keep comparing their partners to mostly fake idols on social platforms.”

[Read: The happiness trinity]

According to Evans, several trends are driving this global decline in coupling. Smartphones and social media may have narrowed many young people’s lives, pinning them to their couches and bedrooms. But they’ve also opened women’s minds to the possibility of professional and personal development. When men fail to support their dreams, relationships fail to flourish, and the sexes drift apart.

If I had to sum up this big messy story in a sentence, it would be this: Coupling is declining around the world, as women’s expectations rise and lower-income men’s fortunes fall; this combination is subverting the traditional role of straight marriage, in which men are seen as necessary for the economic insurance of their family.

So why does all this matter? Two of the more urgent sociological narratives of this moment are declining fertility and rising unhappiness. The relationship recession makes contact with both. First, marriage and fertility are tightly interconnected. Unsurprisingly, one of the strongest predictors of declining fertility around the world is declining coupling rates, as Burn-Murdoch has written. Second, marriage is strongly associated with happiness. According to General Social Survey data, Americans’ self-described life satisfaction has been decreasing for decades. In a 2023 analysis of the GSS data, the University of Chicago economist Sam Peltzman concluded that marriage was more correlated with this measure of happiness than any other variable he considered, including income. (As Stone would rush to point out here, marriage itself is correlated with income.)

The social crisis of our time is not just that Americans are more socially isolated than ever, but also that social isolation is rising alongside romantic isolation, as the economic and cultural trajectories of men and women move in opposite directions. And, perhaps most troubling, the Americans with the least financial wealth also seem to have the least “social wealth,” so to speak. It is the poor, who might especially need the support of friends and partners, who have the fewest close friends and the fewest long-term partners. Money might not buy happiness, but it can buy the things that buy happiness.

This article was featured in the One Story to Read Today newsletter. Sign up for it here.

When colleges began announcing the makeup of their incoming freshman classes last year—the first admissions cycle since the Supreme Court outlawed affirmative action—there seemed to have been some kind of mistake. The Court’s ruling in Students for Fair Admissions v. Harvard had been almost universally expected to produce big changes. Elite universities warned of a return to diversity levels not seen since the early 1960s, when their college classes had only a handful of Black students.

And yet, when the numbers came in, several of the most selective colleges in the country reported the opposite results. Yale, Dartmouth, Northwestern, the University of Virginia, Wesleyan, Williams, and Bowdoin all ended up enrolling more Black or Latino students, or both. Princeton and Duke appear to have kept their demographics basically stable.

These surprising results raise two competing possibilities. One is that top universities can preserve racial diversity without taking race directly into account in admissions. The other, favored by the coalition that successfully challenged affirmative action in court, is that at least some of the schools are simply ignoring the Supreme Court’s ruling—that they are, in other words, cheating. Finding out the truth will likely require litigation that could drag on for years. Although affirmative action was outlawed in 2023, the war over the use of race in college admissions is far from over.

History strongly suggested that the end of affirmative action would be disastrous for diversity in elite higher education. (Most American colleges accept most applicants and therefore didn’t use affirmative action in the first place.) In the states that had already banned the practice for public universities, the share of Black and Latino students enrolled at the most selective flagship campuses immediately plummeted. At UC Berkeley, for example, underrepresented minorities made up 22 percent of the freshman class in 1997. In 1998, after California passed its affirmative-action ban, that number fell to 11 percent. Many of these schools eventually saw a partial rebound, but not enough to restore their previous demographic balance.

Something similar happened at many selective schools in the aftermath of the Supreme Court’s 2023 ruling. At Harvard and MIT, for example, Black enrollment fell by more than 28 and 60 percent, respectively, compared with the average of the two years prior to the Court’s decision. But quite a few institutions defied expectations. At Yale, Black and Latino enrollment increased, while Asian American enrollment fell by 16 percent compared with recent years. Northwestern similarly saw its Black and Latino populations increase by more than 10 percent, while Asian and white enrollment declined. (In Students for Fair Admissions, the Court had found that Harvard’s race-conscious admissions policies discriminated against Asian applicants.)

[Rose Horowitch: The perverse consequences of tuition-free medical school]

Figuring out how this happened is not easy. Universities have always been cagey about how they choose to admit students; the secrecy ostensibly prevents students from trying to game the process. (It also prevents embarrassment: When details have come out, usually through litigation, they have typically not been flattering.) Now, with elite-college admissions under more scrutiny than usual, they’re even more wary of saying too much. When I asked universities for further details about their response to the ruling, Dartmouth, Bowdoin, and Williams declined to comment, Yale and Northwestern pointed me toward their vague public statements, and a Princeton spokesperson said that “now race plays no role in admissions decisions.” Duke did not reply to requests for comment.

The information gap has led outside observers to piece together theories with the data they do have. One possibility is that universities such as Yale and Princeton are taking advantage of some wiggle room in the Supreme Court’s ruling. “Nothing in this opinion should be construed as prohibiting universities from considering an applicant’s discussion of how race affected his or her life, be it through discrimination, inspiration, or otherwise,” Chief Justice John Roberts wrote in his majority opinion. This seemed to provide an indirect way to preserve race-consciousness in admissions. “It’s still legal to pursue diversity,” Sonja Starr, a law professor at the University of Chicago, told me. Her research shows that 43 of the 65 top-ranked universities have essay prompts that ask applicants about their identity or adversity; eight made the addition after the Court’s decision.

Another theory is that universities have figured out how to indirectly preserve racial diversity by focusing on socioeconomic status rather than race itself. In 2024, Yale’s admissions department began factoring in data from the Opportunity Atlas, a project run by researchers at Harvard and the U.S. Census Bureau that measures the upward mobility of children who grew up in a given neighborhood. It also increased recruitment and outreach in low-income areas. Similarly, Princeton announced that it would aim to increase its share of students who are eligible for financial aid. “In the changed legal environment, the University’s greatest opportunity to attract diverse talent pertains to socioeconomic diversity,” a committee designed to review race-neutral admissions policies at the college wrote.

Some evidence supports the “socioeconomics, not race” theory. Dartmouth announced that it had increased its share of low-income students eligible for federal Pell grants by five percentage points. Yale has said that last year’s incoming freshman class would have the greatest share of first-generation and low-income students in the university’s history. Richard Kahlenberg, a longtime proponent of class-based affirmative action who testified on behalf of the plaintiffs challenging Harvard’s admissions policies, told me that, by increasing economic diversity as a proxy for race, elite colleges have brought in the low-income students of color whom purely race-based affirmative action had long allowed them to overlook. (In recent years, almost three-quarters of the Black and Hispanic students at Harvard came from the wealthiest 20 percent of those populations nationally.) “While universities had been claiming that racial preferences were the only way they could create racial diversity, in fact, if we assume good faith on the part of the universities, they have found ways to achieve racial diversity without racial preferences,” Kahlenberg said.

[Richard Kahlenberg: The affirmative action that colleges really need]

If we assume good faith—that’s a big caveat. Not everyone is prepared to give universities the benefit of the doubt. Edward Blum, the president of Students for Fair Admissions, the plaintiff in the case that ended affirmative action, has already accused Yale, Princeton, and Duke of cheating. And Richard Sander, a law professor at UCLA and a critic of affirmative action, said that if a university’s Black enrollment numbers are still above 10 percent, “then I don’t think there’s any question that they’re engaged in illegal use of preferences.”

The skeptics’ best evidence is the fact that the universities accused of breaking the rules haven’t fully explained how they got their results. Yale, for example, has touted its use of the Opportunity Atlas, but hasn’t shared how it factors information from the tool into admissions decisions. Before the Court’s ruling, a Black student was four times more likely to get into Harvard than a white student with comparable scores, and a Latino applicant about twice as likely.

To keep numbers stable, race-neutral alternatives would have to provide a comparable boost. According to simulations presented to the Supreme Court, universities would have to eliminate legacy and donor preferences and slightly lower their average SAT scores to keep demographics constant without considering race. (In oral arguments, one lawyer compared the change in test scores to moving “from Harvard to Dartmouth.”) With minor exceptions, selective universities have given no indication that they’ve made either of those changes.

Even the data that exist are not totally straightforward to interpret. Some universities have reported an uptick in the percentage of students who chose not to report their race in their application. If that group skews white and Asian, as research suggests it does, then the reported share of Black and Latino students could be artificially inflated. And then there’s the question of how many students choose to accept a university’s offer of admission, which schools have little control over. Wesleyan, for example, accepted fewer Black applicants than it had in prior years, Michael Roth, the university’s president, told me. But a larger share chose to matriculate—possibly, Roth said, because even-more-selective schools had rejected them. The University of Virginia similarly had an unusually high yield among Black students, according to Greg Roberts, its dean of admissions. He couldn’t tell whether this was thanks to the school’s outreach efforts or just a coincidence. “I think what we’re doing is important, but to the extent it will consistently impact what the class looks like, I have no idea,” he told me. (Both Roth and Roberts, the only university administrators who agreed to be interviewed for this article, assured me that their institutions had obeyed the Court’s ruling.)

None of those alternative explanations is likely to sway the people who are convinced the schools cheated. With Donald Trump back in office, colleges that don’t see a meaningful uptick in Asian enrollees will likely face civil-rights investigations, says Josh Dunn, a law professor at the University of Tennessee at Knoxville. “If everything ends up looking exactly like it did prior to SFFA,” he told me, then the courts will “probably think that the schools were not trying to comply in good faith.”

Blum, the head of Students for Fair Admissions, has already threatened to sue Yale, Princeton, and Duke if they don’t release numbers proving to his satisfaction that they have complied with the law. (Blum declined to be interviewed for this article.) A new lawsuit could force universities to turn over their admissions data, which should reveal what’s really going on. It could also invite the Court to weigh in on new questions, including the legality of race-neutral alternatives to affirmative action that are adopted with racial diversity in mind. A resolution to any of these issues would take years to arrive.

In many ways, the endless fight over affirmative action is a proxy for the battle over what uber-selective universities are for. Institutions such as Harvard and Yale have long been torn between conflicting aims: on the one hand, creating the next generation of leaders out of the most accomplished applicants; on the other, serving as engines of social mobility for promising students with few opportunities. It will take much more than the legal demise of affirmative action to put that debate to rest.