For the majority of American voters, who picked a presidential candidate other than Donald Trump, the race between two responses — “How?” and “Why?” — remains too close to call well after Trump won the Electoral College and the American presidency. Those questions come from a core concern. It’s not just that The One They Wanted lost, or that The One They Didn’t Want won. It’s not even entirely fear for the next four years, or the half-life of the fallout from Trump’s time in office. It’s all of those things, but it goes deeper than that. In winning the way he did, Trump shook our faith in something fundamental: our ability to predict how our fellow citizens feel. And that’s more unsettling than the prospect of a single Trump term.
As you’ve no doubt read in a pop-science piece somewhere, our species is poorly equipped to process unpredictability. We construct rickety narrative trellises to prop up random events and we worry about flashy, infrequent threats (being blown up by terrorists) while discounting the mundane dangers that are much more likely to kill us (the car we drive to work or the meals we make for dinner). In the past, these thought processes protected us from predators. Now they protect us from from feeling out of control. Election night punctured our fragile firewalls as quickly as Trump took down Clinton’s. The results preyed on our anxiety about unpredictable outcomes in four different ways, each of which whittled away some of our certainty about the way the world works.
First, the polls: the big betrayal, to those who’d come to depend on the models’ lopsided margins as a source of comfort. We watched the numbers, adjusted and raw, rise and fall for months — sometimes slowly and sometimes quickly, depending on the model, the month, and the latest sticky sound bite, but always in a way that produced the same forecast: our first female president. The credible forecasters were so closely clustered on the left side of the spectrum that even the FiveThirtyEight model, which drew fire for being the most slanted toward Trump, had his chances at worse than 30 percent. Collectively, the polls as of Tuesday morning put her ahead by between 3 and 4 percentage points in the popular vote; she looks likely to finish ahead by less than 1 percentage point. Adding to the postelection uncertainty, Tuesday’s Clinton-leaning exit polls were wildly misleading, and her patterns of under- and overperformance fluctuated from state to state.
A probabilistic forecast isn’t necessarily “wrong” when an unlikely outcome occurs. Even a model that gave Clinton a 99 percent probability of victory allows for one anomalous spin of the wheel. It’s also possible, though, that the wheel was tilted toward Trump in a way that the polling (and, thus, the models) missed. The popularization of public political forecasting coincided with a pair of predictable presidential races in 2008 and 2012; maybe that made us too sanguine about the state of the latest race. For now, we can’t heap the whole blame on a particular part of the electorate (although whites without college degrees are the leading suspects), but in the days to come, forensic analysis will suss out the sources of Trump’s appeal, pinpointing the places where turnout and party loyalty were higher or lower than we were led to expect. Even when we get answers, though, they might come too late to restore the public’s trust in data-driven projections. It’s easy to imagine future forecasting attempts being greeted by frequent refrains of “But what about Trump?”
For those who think of predictions in binary terms — Clinton or Trump — the probabilistic explanation isn’t satisfying. The models aren’t based on what a computer projected people to do; they’re based on what people reported that they intended to do (or, in the case of exit polls, what they claimed to have done). Why did surveys of so many respondents steer us wrong? Maybe the polls were conducted poorly; maybe Trump voters were reluctant to disclose their stance. Regardless, after all the regressions and poll ratings, the most accurate forecaster may have turned out to be Trump, who in August dubbed himself “Mr. Brexit.” On the heels of his election, the poll-defying results from the U.K.’s EU referendum, and several other recent, high-profile polling failures, we’re left to wonder whether we’re in the dark. If Trump’s opponents had seen his victory coming, they might have been mentally prepared. Instead, he won with disconcerting surprise, like a movie villain who puts a bullet in Bond instead of helping him stall for time. So much for the ground game.
In addition to compromising our confidence in predictions, the election eroded our belief in how well direct democracy works. “What’s the Electrical College?” I heard someone ask, seriously, on Tuesday night. It’s easy to mock the malapropism, but many of the voters who know it’s actually “Electoral College” still don’t know why we need it. Maybe we won’t, sometime soon, but in 2016 we’re stuck with a system that makes elections more arbitrary as well as more difficult to call. From our first student-council experience, we’re conditioned to think that the primary goal of a race is to get the most votes. But for the second time in five elections, a candidate appears to have won the presidency despite getting fewer votes than the loser, abrading whatever slight sense of control we derive from casting a vote.
And then there’s the stigma of supporting Trump, the candidate the polling promised was the least-liked on record. Of all the weak contenders, the spoils went to the one who sunk lowest and yelled loudest; who seemed most self-absorbed and ill-informed; who showed, through his words and his deeds, his disregard for (or hostility toward) anyone who wasn’t like him. Never had a major party’s candidate given Americans so many reasons to root against an underdog; never had one survived (if not benefited from) such a lack of qualifications. To a Trump opponent, it’s almost inconceivable that a reasonable person could study his reality-star record and still vote to make him the most influential American; it’s even more galling to know that some voters filled in Trump’s bubble without paying attention to his past. Presidents might not matter as much as we think they do, but in this case, the choice of president does.
In our first few years of life, we develop the ability to anticipate and empathize with the behavior of others. We depend on that predictability, because we have no choice but to trust strangers: when we stand on a subway platform; when we drive toward an oncoming car in a neighboring lane; when we wait for the light and cross at the corner. Both Clinton and Trump partisans believe that by voting the way they did, many millions of others just did something self-destructive and dangerous. That disillusionment allows another layer of uncertainty to set in: Our instincts were no better than the polls at predicting what people would do.
Lastly, there’s the uncertainty surrounding what’s in store. As the polls started closing, the results started rolling in, and Trump’s odds started increasing, a bewildered diner at my table wondered aloud, “What would Trump look like as president? Really, what would he look like?” We know what he’d literally look like — ochre complexion, Fallon-tousled hair, ill-fitting suit, too-long tie — but his distinctive uniform camouflages his nebulous policy positions. Ironically, Trump’s amorphousness on the issues may have made him more electable, but it’s also increased the perceived risk of placing huge power in Trump’s tiny hands. The odds of a catastrophe are extremely low; then again, so were the odds that Trump would win. And after Tuesday’s reminder that we know less than we thought, it seems more plausible that if a disaster were to develop, we wouldn’t see it coming — which in turn sends the message that it’s never safe to stop worrying.
Stripped of reassuring assumptions, we’re left with a toxic mix of emotions. There’s the sadness that Barack Obama’s presidency, which began with the oft-fulfilled promise of positive change, will end with two months of lame-duck dread, culminating in a handoff to a temperamental and intellectual Bizarro Barack who’ll attempt to repeal the outgoing president’s policies. There’s the dismay that some minorities might justifiably feel less welcome in the country than they did before. There’s the frustration of facing an indefinite period of suspended progress and the awareness that there’s no telling how long it will take for a more electable woman to end the office’s 227-year-and-counting XY-chromosome streak. There’s the shame of having the face we present to the world be one on which most voters swiped left.
I wrote this on an overnight flight from Phoenix to Newark. Internet was expensive, so I stayed offline, happy to live for a few more hours in a country that, as far as I knew, hadn’t yet picked a president. The ostrich approach worked until a passenger sitting near me announced, “Trump’s at 276,” and another confirmed to a flight attendant that Hillary had conceded. “You know what I want to know?” the attendant asked. “When was the last time he even voted?” (Actually, 2014.)
If we took the direct route, we flew over 10 states, six red and four blue, each bout of turbulence and lighting of the seat belt sign seeming like on-the-nose symbolism. Even in broad daylight, red and blue boundaries aren’t apparent from 5 miles up, and at night, every state looks like featureless blackness broken up by islands of light. The larger islands, I know, were more likely to have voted for Clinton, but each pinprick I saw represented someone whose head could have concealed any emotion, from happiness to heartbreak. I usually find cross-country flights affirming: The scenery scrolls by, and I think, “This is all us.” Maybe the feeling will fade, but on this flight, the landscape felt foreign.
Beginning next January, Donald Trump will be president for the foreseeable future. And because of how we arrived there, that future now seems more unknowable, and more frightening. On the bright side, there’s some slim chance that he’ll make a passable president; sometimes, low probabilities pay off.