Monumentally Wrong
A year ago in this magazine, a seasoned political journalist had a message for reporters covering the presidential primaries: We blew it.

Here we are again.

When future Amherst students hearken to the Dark Ages of 2016—back when ancient architectural artifacts apparently known as the “Social Dorms” met the wrecking ball—only one thing should be clear about the presidential campaign: the past is not always prologue.

So, future students, beware when even the most learned in your midst warn you of being too bedazzled by the new and shiny. Back in 2016, Donald Trump proved new, shiny and totally confounding to the most fundamental assumptions of the day.

Where does one start with how wrongheaded the media (this author included) and political establishment were about what it takes to win the White House?

Did you wager on Wisconsin Gov. Scott Walker or Ohio Gov. John Kasich upending a well-funded Jeb Bush juggernaut to procure the Republican nomination? Did you assume Trump was dead early after doubting John McCain’s heroism and claiming thousands of New Jersey Muslims cheered the destruction of 9/11? Was Hillary Clinton a lock cinch for you because of the unalterable imperatives of the electoral map?

I must raise my own hand to all those 2016 prospects.

For starters, a cranky Brooklyn-bred Vermont socialist, Bernie Sanders, could not be a serious challenger to Clinton, the presumptive Democratic nominee. Right? And the Republican Party establishment would carry the likes of Bush, Kasich, Walker, Marco Rubio, Chris Christie, even aggressively dour Ted Cruz, to victory.

But Trump, the political neophyte, self-aggrandizing entrepreneur and TV celebrity who grappled daily with the truth? No way.

There were, after all, the many basic tenets held to with biblical certitude about modern elections. They dominated the derision toward Sanders and Trump. They included: The candidate who raises the most money wins; changing demographics (the African-American and, especially, the Latino vote) would be destiny for Clinton; TV ads were crucial even in a digital age; party unity was essential and the GOP lacked it; and staying on message is incumbent.

There’s more: Debates would be impactful this time and frighteningly unmask Trump’s superficiality; rhetorical self-discipline is essential; avoidance of big mistakes is crucial; policy flip-flops are dangerous; and there is the ironclad necessity for a sophisticated and vast field operation supplemented by digital targeting of prospective voters, à la Obama in 2008 and 2012.

Most everything proved utterly wrong.

The consistent strategic and tactical errors of assessment were up there with those of George Custer at Little Big Horn, the design flaws of the RMS Titanic, Decca Records spurning a Liverpool band called The Beatles and Julius Caesar rejecting his wife’s nightmare-inspired admonition to, please, not head off to the Roman Senate.

Just take our growing homage to analytics, a favorite topic of Steve Edwards ’93, a longtime journalist who is now executive director of the Institute of Politics at the University of Chicago. This campaign, he notes, was “supposed to represent the triumph of data in modern presidential campaigning” since predictive methods were lauded with powerful election victories of Obama and predecessor George W. Bush.

Remember? Data teams became stars, with many then spinning myths into
substantial riches in the private sector. I’ve met multiple individuals acclaimed as digital geniuses for “spearheading” one Obama campaign or the other, leaving me to wonder how many spears were deployed.

Edwards, too, assumed Clinton would win, despite the “change election” dynamic and the challenge of succeeding a two-term president of one’s own party. But hallowed data models, like those of data superstar Nate Silver, failed miserably in predicting Trump’s triumph.

“In that sense, I think one of the big takeaways of 2016 is that the best predictive analytics tools in the world aren’t a substitute for talking to real people,” says Edwards. “That’s true whether you’re a pollster, a journalist, a pundit or a campaign manager.”

But polling mania persists. It’s an easy fallback, especially for media looking for faux daily news and punditry. Our ignorance, as both consumers and practitioners, is ample, notes Tanya Leise, associate professor of mathematics and statistics at Amherst.

“My main concern with election polling and prediction is that most consumers of these efforts don’t really understand the uncertainty involved or how to quantify it,” says Leise, who taught a fall 2016 course on the math of elections (see page 27), including whether systems such as plurality, Borda count, Hare/IRV and beatpath/CSSD “satisfy mathematical properties like monotonicity, Condorcet winner criterion, majority criterion, etc.”

Got it? Most journalists certainly do not.

And while Leise’s class did not study issues such as prediction via statistical models, she knows that “humans intrinsically tend to be poor at making probabilistic calculations, and we like to have clear answers. Unfortunately, both polling and prediction attempt to identify trends occurring in very complex and dynamic situations, and so will naturally have a great deal of uncertainty involved, implying that predictions will have to be probabilistic in nature.”

Models that aim at forecasting election results “yield probabilities of various outcomes. For one-time events like this past U.S. presidential election, that means we can’t definitively say a model was wrong (unless it gave a 0 percent probability of Trump winning). A low probability doesn’t mean it can’t happen; sometimes dynamic events converge just so, or the ‘true values’ of polled information turn out to be at the edges of the margins of error. If we roll a pair of dice, saying there’s a 97 percent chance of rolling anything other than snake eyes, it could still happen that we do roll snake eyes.”

Of course, there are a few folks who felt earlier than most that Trump might be a winner. They are exceptions that prove a rule of relentlessly misguided conventional wisdom.

“About a month before the election, before the revelations of Trump boasting of sexually assaulting women and [FBI director James] Comey’s intervention in the election, as well as the Russian/WikiLeaks revelations, I started telling friends and colleagues that I thought Trump would win. Austin Sarat will verify this if you ask him!” says Thomas Dumm, the William H. Hastie ’25 Professor of Political Science.

“My reasoning was simple, perhaps simplistic. When my friends would be incredulous, I would give them the following quiz: Aside from the fact that she isn’t Trump, why are you voting for Hillary Clinton?”

e adds, “While some of them could, after hemming and hawing, come up with something like, ‘She is a woman,’ ‘She favors civil rights,’ or other fairly empty answers—empty in the sense that alone they didn’t mark any substantive claims to the future policy of the U.S.—no one immediately had an answer to the question.”

While Dumm was sending out his yellow warning light, he was in a clear minority of academics and the media. The overriding consensus even as polls began closing in the East and Midwest on Nov. 8 was that Trump’s electoral hill was too steep to climb. Media pundit after media pundit reiterated the tortuous, even impossible path for Trump: He’d have to win Florida, then Michigan and Pennsylvania or Ohio. It just wouldn’t happen. The press echo chamber has never been louder.

We were taken to the cleaners in abysmal analyses of key states. Even when curious returns started coming in from rural locales, they were first seen as anomalies. “Is it possible?” I asked friends in big-time newsrooms. It was very late before Trump’s lack of a concession speech was seen as having potential justification. The subsequent sense of bewilderment has never been as vivid. The past had to be prologue. Demographics ordained a Clinton victory. Trump made too many too skittish. Right?

U.S. Rep. Tom Davis ’71, a Virginia Republican whose wife ran for lieutenant governor in 2013, discerned years ago that the GOP base had moved dramatically away from his moderate essence. He backed John Kasich. Under vaguely normal circumstances, Trump “would not have gotten to first base,” Davis says. But social and economic disruption over the Obama years inspired anxiety. Even if one saw this as a “change election,” Trump did seem a bit much for a long time.

But “he got discipline at the end and squeaked it out. He peaked at the right time,” Davis says. Like many ultimate Trump voters, Davis was not terribly enthusiastic. But he didn’t want Clinton.

Davis is board chairman at George Mason University and has run many campaigns. He concedes how many of us “live in bubbles”—an unavoidably self-incriminating remark, since Davis, too, failed to see this coming. The people who were caught are of varied cohorts, not just alleged cognoscente in New York-Washington-Boston. Too many people are too isolated in self-justifying universes.

They don’t get out. The media is the easiest target—once again, it preferred the simpler task of covering the candidates, not the country—but it’s not alone. How many people ventured into rural America, even longtime Democratic strongholds, to talk and ultimately perceive a deep unease that manifested itself in the unexpected pro-Trump turnout there?

“Most of us probably never watched The Apprentice,” Davis says.

Well, I watched it and still blew it. If only I could eliminate from digital view my previous declarations as easily as we bid farewell to the Social Dorms. 

James Warren ’74 is chief media writer for the Poynter Institute for Media Studies and a national columnist for U.S. News & World Report. He’s the former managing editor of the Chicago Tribune.