The Church of Averageness: 'Polls Say' (Hate Inc.)

This is a free chapter from “Hate Inc.” (formerly “The Fairway”). If you’re already a subscriber (thank you!), please consider sharing it. If not, you can subscribe to get all chapters in your inbox as they’re published.

Donald Trump was the least “electable” presidential candidate in modern history. He probably beat out even Lyndon Larouche on this score. 

Trump grievously offended nearly every voting demographic. He’d teed off on women, Latinos, Muslims, the disabled, “the blacks,” veterans, and Asians (“We want deal!” he cracked, about the Chinese).

He retweeted, about Iowans, before the Iowa Caucus, the line, “Too much #Monsanto in the #corn creates issues in the brain?” This was after being confronted by a poll showing him trailing Ben Carson in upcoming caucuses. Repeat: he ripped Iowa, Monsanto, and corn before the Iowa caucuses.

Additionally, Trump violated every idea we had about what a presidential candidate looked, acted, and sounded like. He threw water, bragged about his dong size, ranted about women’s periods, and while doing so, didn’t check other key “electability” boxes.

He had no “ground game,” a characteristic normally cited as a crucial factor. He was also an adulterer and ignorant of the Bible, running in a primary whose constituents supposedly treasured religion.

By every conceivable standard of conventional wisdom, Trump had no shot. As noted before, even data journalists laughed at the notion that he could win.*

Nate Silver, the former baseball stats guru turned National Oracle™ (as Gizmodo called him), said Trump had a better shot of “cameoing in another Home Alone movie with Macaulay Culkin (or playing in the NBA Finals) than winning the Republican nomination.”

But he won anyway. This should have proved “electability” was a crock, and killed it forever as a form of campaign analysis. 

It did not. It lives on, as journalism’s version of junk forensics. It’s like the infamous “comparative bullet-lead analysis” – a made-up trick, about as scientific as chicken-bone divination, that nonetheless routinely impacts devastating real-world decisions. 

“Electability” often amounts to pundits saying that because a thing just happened, it will never stop happening.

Take a close race, like the 2000 presidential election. In it, independents broke late for George W. Bush, snapping a near-statistical tie.

An interesting statistical observation would have been more in line with what Chomsky said about that race. His take was that in a sample size that enormous, a tie would only be expected in one situation: if people were voting for something random, like the presidency of Mars.

The amazing closeness of American elections has never made sense. In a country in which 10% of the population owns 90% of the wealth, you’d expect the very rich to be a permanent electoral minority. That it doesn’t work out that way is odd. But this is not the kind of observation pundits tend to make. Instead, we peck around the surface.

The take from 2000 was about how independents behave when things get close. Pundits cited a statistical law: in tight races, undecideds tend to break for the challenger.

This for a time was called the “incumbent rule.”

Four years later, the “incumbent rule” was broken out to raise concern about George W. Bush’s seemingly stagnant support levels heading into Election Day 2004, against challenger John Kerry.

When Bush won, suddenly everyone was wondering what happened to that incumbent rule.

So some cited a Democratic pollster, Mark Mellman, who wrote: “We simply do not defeat an incumbent President in wartime.”

So that maxim was added to the other. Both became conventional wisdom:

Tight races tend to break away from the incumbent, unless it’s in wartime, in which case they don’t.

Reporters built up a stack of these “laws of campaigning.” We became alchemists in big conical hats, sorting through giant tomes at the start of each race, to see if paths to victory for each candidate could be found in our mazes of rules.

The trick was using polls to convince voters to interpret political news through someone else’s eyes, instead of their own brains. You may like the policies of candidate X better, but “polls say” (this use of the passive voice is key) you should vote candidate Y, if you want to win the election.

But “Polls say” is often just “we say,” disguised – in the same way a man-on-the-street quote is often just the first person found who agreed with a point the reporter wanted to make.

In elections past, “Polls said” an “electable” candidate was someone capable of “crossing the aisle,” a “fiscal conservative” (whatever that meant), not “soft on defense” and possessing of certain personal characteristics: married, with kids, heterosexual, tall (seriously), presentable, religious, and preferably with military experience.

Both parties and their main donors consistently threw, and throw, their money behind candidates who check all of these boxes. This person becomes the presumptive frontrunner. Campaign reporters would then trail along – I’ve watched this many times – and prod would-be voters at events to comment on the candidate’s superior “electability.”

This was particularly an issue with John Kerry. If you were a reporter following Kerry, you felt like you’d died and woken up in a vat of boiling grease. The man was pure distilled boredom. He had no clue why he was running for president.

The only thing Kerry seemed to enjoy on the campaign was “orange baseball,” a game in which he’d roll an orange from the front section of his campaign plane down the aisle into the press section.

The vets on the plane explained this was an old tradition (apparently Nancy Reagan was really into it as well). But Kerry bowled more oranges than any candidate they had ever seen.

If we were taking off, Kerry would lean his huge head into the aisle, flash a dazed smile, and drop an orange on the carpet. As the plane accelerated upward, the fruit would fly down the aisle, gaining speed throughout, and smack the piles of camera equipment at the back of the plane, if it didn’t first bounce into the face of a sleeping reporter or bash an unsuspecting stewardess in the kitchen. Kerry would smile equally at each of these outcomes.

The “fun” soon became weird. The female reporters in particular found the whole thing obnoxious. Up front, you could plainly see that Kerry’s team was trying to talk to him about stuff, but he was focused on that orange. On one flight we had two going at once.

What was up with this dude? All we knew about Kerry was that he was supposedly more “electable” than Edwards (who was too “angry” on class issues) and Dean (who was “too liberal”). And this was, after all, the “electability” election. In 2004, the buzz was, nothing was more important than electability.

How did we know that? Because everyone in the plane said so. There were literally thousands of articles about Kerry and “electability” that year.

Matt Bai of the New York Times later summed up 2004 as follows: “In this year's campaign, electability became the issue itself.”

After playing the orange game on Kerry’s flights, reporters would jump out in search of fodder for “electability” stories. At events, which after all were filled with Kerry supporters, they’d ask questions like, “Do you think Kerry can beat Bush?”

“Uh, yes,” the person would say. “I think he can beat Bush.”

Next morning, you’d see the story:

John Q. Dinglehat of Hologram, New Hampshire wants to see a Democrat in the White House – and thinks John Kerry is just the man to make it happen. ‘I think he can beat Bush,’ Dinglehat says.

Pollsters and pundits alike framed Kerry as the “beat Bush” candidate. “Seen as the Best Candidate to Beat Bush, Kerry primed for a N.H. Victory,” the Gallup service wrote, just before the primary.

At that and other primaries that year, the National Election Pool exit poll questionnaire asked voters if they cared more about having a nominee with whom they agreed on policy, or one who “can beat George Bush.”

And just like that, electability really did become the driving issue of the race, when it never had been before.

In 2000, New Hampshire voters answering exit surveys had listed “best chance to win in November” as their primary reason for choosing a candidate just seven percent of the time.

By 2004 they were listing it as the primary reason between three and five times as often. In Iowa, an amazing 50% of voters listed it as their chief concern. We were asking the fuck out of that question.

Soon enough, as Kerry racked up primary wins, our own language started to bounce back to us. This is from the Washington Post:

Patricia Coan of Fairfax, a retired medical practice manager, said: “I voted for Kerry for the same reason most Democrats are voting for him – he can beat George Bush.”

The phenomenon worked in reverse with other candidates. Reporters would pester people: “Will you vote for Edwards even if someone else has a better chance to win?”

Or: “Does it bother you that Kucinich has no chance to win the election?” I’d hear reporters laying this on people in August and September of 2003, months before the New Hampshire primary.

Dean actually got the following question in an online Q&A in November of 2003: “How do you address the common criticism: ‘I like Dean's policies, but I don’t think he can beat Bush?’”

The whole “electability” question usually implies a) there’s a candidate in this field who’s most likely to win, and b) there’s a candidate who appeals to you on a policy level, and c) those candidates are not the same person.

People to this day believe this is the case. Voters over generations have been trained to think the politician who represents their views is unlikely to be “electable.”

Most people are terrified of throwing their vote away, so they’ll steer way clear of any candidate the press tells them has no chance. Particularly when the incumbent is odious, voters won’t vote their own interests and conscience. They actually think it’s their civic duty not to.

The trick works best with political minority groups, who’ve been trained to vote according to how they’re told a larger plurality thinks. Until pretty recently, if you were nonwhite, female, single, childless, or gay, you were typically told you had to choose between a slew of straight white candidates who “polls said” had an actual chance.

That dynamic is still true if you belong to any non-traditional political persuasion. If you’re an anarchist, socialist, a “populist” (this term can mean almost anything), a nationalist, a Green, even a libertarian (the Ron Paul kind, not the Peter Thiel/Koch Brothers kind), you’ll hear that “polls say” people are not ready to elect your type of person.

This effect is so powerful that it caused Barack Obama to underperform with black voters early in the 2008 race. Obama did not get the early backing of a lot of black churches, and leading African-American Democrats like John Lewis initially steered clear. 

This was not because Obama was a poor candidate, but precisely because he was a “plausible” one. African-American voters, the buzz went, were genuinely afraid he would win the nomination, then lose to a Republican.

“They didn’t know him, a), and, b), they thought it was a long shot,” said Jesse Jackson. “Black voters are comparatively conservative and practical.”

When Obama made enough gains that year that black voters started to move in his direction, suddenly there was a new party line: Latino voters were the new “firewall” for Hillary Clinton.

It goes without saying that language about “firewalls” is crazy and insulting. It implies the nomination is the property of the presumptive frontrunner, and challengers are destructive forces. In this case it was openly argued before the primary in California (where fires are sort of an issue) that Clinton’s best hope was the historical Latino distrust of black candidates.

“The Hispanic voter – and I want to say this very carefully – has not shown a lot of willingness to support black candidates,” a Clinton pollster named Sergio Bendixen told the New Yorker. Robert Novak later summed up: “Clinton Gambles With Latino Firewall.”

The “polls say” trick also works, sadly, with labor. Every year, even in the primaries, unions endorse candidates with poor records on labor, because they buy the “electability” pitch.

Sometimes small locals will back the obvious pro-union candidate, while national leadership will give the big endorsement to the more ambiguous frontrunner. This was the case with Clinton over Sanders in 2015–2016.

But big labor voting against big labor is a regular theme. I watched this at an AFL-CIO conference in Whitfield, New Hampshire in 2003.

At the event, each Democratic candidate made a plea for the labor vote. Two were longtime union favorites: Richard Gephardt of Missouri and Ohio’s Dennis Kucinich. These politicians checked every single organized labor wish list. Gephardt had a 100% rating with the national AFL-CIO, as did Kucinich, who among other things called for a repeal of NAFTA and the WTO.

But the group in the end decided to back Kerry, who had voted for Most Favored Nation status with China, was a staunch NAFTA defender, and strongly believed in global trade policies.

“We need a seat at the table,” is what one of the union men told me, implying that it was better to back a weak-on-labor Democrat with a shot than a good one with none.

If labor won’t back a lifetime labor advocate like Dick Gephardt, surely college kids would vote for Kucinich?

After all, Kucinich was the only candidate who treated college students like grownups and embraced idealistic policies like a Department of Peace. While other campaigns tried to win over “youth” by passing out T-shirts with cutesy names (“Deanie Babies” and “Liebermaniacs”) or giving away free hot dogs, Kucinich went the other way.

At the University of New Hampshire in Durham, he refused to dumb down his speech and quoted the likes of Jung, Barbara Marx Hubbard, Thomas Berry, and the humanist sociologist Morris Berman. 

In a preview of issues that would become extremely popular among Democrats years later, he said his campaign was about changing the priorities of the country away from military spending and foreign intervention, and toward greater investment in education and health care. He sat down and opened up the floor to discussion, and the “stump speech” became a sit-in, where people talked about all sorts of things, including their personal lives, depression, drugs, etc.

Kucinich ultimately got a standing ovation (as both he and Gephardt had at the AFL-CIO conference). Afterwards, though, few students said they would even consider voting for him.

“[Kucinich is] everything that I personally would want in a President,” a grad student named David Wilmes told me. “But… it’s going to have to be someone like Kerry or Edwards.” 


“Well,” he said. “It’s probably going to have to be someone who’s tall.”

I told Kucinich about this exchange. He sighed and said that until people learned to vote according to their own beliefs and preferences, politics would be a “mirrored echo chamber, where there’s no coherence.”

This term, “no coherence,” defined modern American politics for a generation, when almost nobody voted for the candidate they personally liked the most.

Middle and lower middle-class Republican voters still endorse tax breaks for billionaires. Labor votes against labor. Inner city minority voters endorse candidates who pledge to lengthen prison sentences and put more cops on the streets. Right-wing voters driving around on Medicaid-funded scooters applaud candidates who rail against the consumers of “free stuff.”

Very few campaign journalists ever raised an eyebrow about this, though, until polls stopped being able to predict it. This is a kinder way of saying that in 2016, voters started to blow off polls, which threw the whole business in disarray.

I hate ripping on Nate Silver. For one thing, I was a big fan of Baseball Prospectus. For another, his essays on polling are always interesting and fun to read. To me he’s been a victim in a larger con.

“Electability” was essentially conventional wisdom in search of scientific recognition. When appeared on the scene, campaign reporters – I heard this – felt their made-up takes were finally being sanctified by data. I remember particularly a debate about Sarah Palin in the later stages of the 2008 race.

In a preview of 2016, reporters were torn. Palin seemed to be the candidate you’d rather have a beer with, but she was also clearly, you know, unfit for office, which bothered a lot of people covering her. How to square the contradiction? They’d normally be celebrating the “beer” candidate, but what to do now?

(This issue somehow didn’t bother the press so much with George W. Bush, which never made sense, but whatever.)

Silver at that time wrote an article explaining that a vice president is by definition assuming office during a crisis. After all, the president has just died. Therefore, experience counts more than “vision.” In sports terms, you’d rather have the major-league ready player than the upside prospect.

“I think Americans can feel sympathy for Sarah Palin, can believe she’s the sort of person they’d want to have a beer with,” he wrote, “and still find her a detriment to McCain’s case for the White House.”

Reporters breathed a sigh of relief. The “beer test” was still a thing, just not in this particular situation! Nobody had to come out and say, “All that stuff we’ve been saying about beer was irrelevant and irresponsible.”

After two election cycles of near-perfect prognostication – he got 99 out of 100 states right in the 2008 and 2012 races – Silver was elevated to God status by campaign reporters. There were stories like “Triumph of the Nerds” that described him as a “big winner” after elections.

The problem was that Silver’s predictions were based on a generation of voter behavior skewed by mountains of our goofball campaign reporting and idiotic conventional wisdom. Should voters ever tune us out, all that data would become meaningless overnight.

This happened in 2016, when Silver outlined a Unified Field Theory of presidential campaign narratives. There were six stages of campaigning, he said, and each portended doom for Donald Trump.

The stages were: Free For All, Heightened Scrutiny, Iowa and New Hampshire, Winnowing, Delegate Accumulation, and Endgame.

Heading into the 2016 election, there were clues that the power of conventional wisdom was waning. A 2015 Pew survey, for instance, showed declining interest in “electability.” I personally was getting that sense less from polls, and more from the increasing frequency of being told to fuck off at campaign events, particularly Republican ones.

The thing was, nobody in the press had any clue what would happen if people stopped listening to our “electability” horseshit. There was no data on this. We were about to find out.

Trump demolished Silver’s “six stages.” He lost Iowa, and even underperforming media expectations there didn’t dent him. He accelerated during the “Winnowing” period. As Silver predicted, the Republican Party did indeed do everything it could to stop him in the Endgame, but if anything that probably helped Trump win the nomination, because Republican voters by then hated the Republican Party.

Official endorsements turned into a negative, not a positive, so having less than 5% of them didn’t hurt Trump. As such, the so-called “establishment primary” was not a factor.

Silver to his great credit ultimately realized that “conventional wisdom” had infected polling. He warned about going too far in the other direction, but his basic take was that “polls may be catering to the conventional wisdom, and becoming worse as a result.” This was in 2017.

He’d already had a dramatic change of view during the 2016 race. By the time election rolled around, he was giving Trump a 29 percent chance of winning, which was higher than just about everyone else.**

There was a big lesson in this for everyone, or there should have been. Politics, despite the fact that it talks about itself as baseball all the time (“inside baseball” is the favored term of people who think they’re playing it), is not baseball. In baseball, batters don’t intentionally strike out because they’re told the pitcher has high strikeout rates.

Data journalism works a check on conventional wisdom. When they combine is when you get problem. The two genres can be as hard to separate as humping dogs. Even after the Trump fiasco, the product of such unholy unions – “electability” – is still running loose.

A summer 2018 piece in the New York Times about the likely Democratic field echoed ancient articles about the likes of Mitt Romney and John Kerry. We were told to be wary of extremists like Elizabeth Warren and Bernie Sanders, and go with something tamer and nearer to the “center”:

For all the evident support for Mr. Sanders’s policy ideas, many in the party are skeptical that a fiery activist in his eighth decade would have broad enough appeal to oust Mr. Trump…

Mr. Sanders’s generational peer, Mr. Biden, 75, is preparing to test a contrasting message this fall… Mr. Biden has struck a gentler chord than Mr. Sanders and Ms. Warren, delivering paeans to bipartisanship and beckoning Democrats to rise above Mr. Trump’s demagogic taunts.

Michael Shearer of the Post argued something similar:

Like several other Senate candidates eyeing 2020, Warren has endorsed a suite of expensive policy proposals that have made some in the party nervous…

For this reason, some Republicans have signaled that they would welcome a Warren run in 2020. Stephen K. Bannon, a former aide to Trump, dismisses Warren as “the weakest candidate the Democrats could put up.”

The easiest way to predict what kinds of “electability” stories you’ll see in an election season is to look at the field of candidates and see which ones have a lot of lobbying and ad money behind them.

Those candidates will be described as electable. Everyone else will get the “polls say” treatment. Be wary of our version of junk science.

* I publicly bet on Trump to win the Republican nomination in August of 2015, but badly misfired in the general election.

** Including me. I went from totally disbelieving in polls to saying, by the end of the race, that Trump “can’t win.”