Winning the Popular Vote While Losing Grip on Reality

Graphic source: Kevin Urmarcher, Kevin Schaul, and Dan Keating, “These Former Obama Strongholds Sealed the Election for Trump,” Washington Post (9 November 2016).

here has been a bit of push back to my last post. A lot of it revolves around this fact: Donald Trump did not win the popular vote. Others point out that if the vote had shifted 1-2% in favor of Clinton, Trump’s stunning electoral sweep would never have happened.

That’s all true. It is also irrelevant.

When I say that large parts of the press and the Democratic establishment have lost grip on reality, I am not saying they are less in touch with ‘the popular will’ than team Trump was. Were you to line up the entire country and rank them by their preference for Mr. Trump or Mrs. Clinton, the median individual (though repulsed by both candidates) would probably lean towards the latter. That is undoubtedly true for the median voter.

It is also utterly irrelevant.

The American presidency is not decided by a popular vote. It has never been decided by a popular vote. The win must come in the electoral college. Hillary Clinton knew this. Her goal was always to win the electoral college. She organized a gargantuan super-PAC network, spent a billion dollars, hired thousands of staffers, mobilized tens of thousands of volunteers, launched one of the biggest get-out-the vote campaigns in this country’s history, and cajoled or inspired thousands of newspapers and media-hands to endorse her in order to win the electoral college. The battle she chose to fight was always about winning a certain alignment of electoral votes.

The problem is not that Clinton lost this battle. The problem is that no one had any idea that the loss was coming. Or that the loss was possible. Or even where the battle would be fought. Clinton, her team, the vast media apparatus that had grown up around it—all were soaking in the same cesspool of self-deceit. The election has shown them all for what they are: an insular network of operators and opinion-makers charmed by their own cleverness and enthralled with their own moral certitude, more comfortable exchanging clever quips and flattering platitudes than confronting the world outside of their carefully constructed echo-chamber.

To put it another way: in this election, narrative building trumped strategy making. Narrative building trumped reporting. It trumped honest assessments of the facts. It trumped everything except Trump himself.

It needs to stop.

Two days ago Michael Cieply gave us an insider’s account of how bad this problem has become, describing his stint at The New York Times:

Historically, the Los Angeles Times, where I worked twice, for instance, was a reporter-driven, bottom-up newspaper. Most editors wanted to know, every day, before the first morning meeting: “What are you hearing? What have you got?” 

It was a shock on arriving at the New York Times in 2004, as the paper’s movie editor, to realize that its editorial dynamic was essentially the reverse. By and large, talented reporters scrambled to match stories with what internally was often called “the narrative.” We were occasionally asked to map a narrative for our various beats a year in advance, square the plan with editors, then generate stories that fit the pre-designated line. 

Reality usually had a way of intervening. But I knew one senior reporter who would play solitaire on his computer in the mornings, waiting for his editors to come through with marching orders. Once, in the Los Angeles bureau, I listened to a visiting National staff reporter tell a contact, more or less: “My editor needs someone to say such-and-such, could you say that?”[1]

That’s just an excerpt. Read the whole thing. Be scandalized by it! Slavish devotion to narrative is not compatible with honest reporting. Moral posturing has no part in candid calculation. If you are serious about gaining and wielding power, then things must change.

I’ll concede that not all of this can be blamed on ideological blinders and class smugness. Secular trends in the media world have played a large part, stripping newspapers of their reporters, centering their content more and more on hot-takes and op-eds. The biases of a digitally centered media are clear. Stories marginal to the election’s end outcome—say, the rise of millennial “alt-right” twitter trolls from the morass of 4chan, PUA forums, etc.—were covered in obsessive detail, while stories central to what actually won Trump the election—say, the complete flip of traditionally left leaning communities in Minnesota(!) and Wisconsin(!)—were ignored. If most of your reporting starts on twitter, well that makes sense. It is easy to see why you might think hashtags and hate filled twitter streams were the real story of this election. Yet the world is bigger than that. The greater part of this week’s voters have never had a twitter account.

This problem is most sickening at the new brand of e-zines which dominate the liberal media scene (e.g., Vox, Salon, Slate. These places don’t do reporting. They rarely even attempt real analysis. The only thing they do well is messaging. Whether the narratives they weave have any connection to reality is, and always has been, a secondary concern to boosting their their candidates and flattering their readers. To explore a comparision I suggested in the last post, place Ezra Klein’s breathless cheerleading for Clinton (“Clinton is an extraordinarily talented politician… winning a process that evolved to showcase stereotypically male traits using a stereotypically female strategy…And it’s working”) earlier this year next to a more sober critique of Clinton and her class written by Thomas Frank a few months earlier.  Frank understood what Klein refused to see. One reveled in the echo dome; the other traveled in the wilderness. One can be trusted to find and tell the truth. The other cannot.

The cesspool must be drained. Flip to Vox now and see the headlines: We just watched Donald Trump try to be President-elect Trump. It took a few tries.” Are these kids serious? Donald Trump is not trying to be President-elect. He is President-elect. The Democratic Party has been decimated. At the presidential level we have just witnessed the party’s largest loss in a generation. At the state level we’ve seen the party weaker than it has ever been in American history. How do the good writers of Vox respond to all this?  With a self-satisfied smirk.

These people must be treated as the pariahs they deserve to be. 

There is more to say, but I’ve said enough for the moment. Instead I’ll close with a few observations Nathan Jurgenson put up on his Tumblr earlier this week. To quote:

It was around 6-to-9pm last night, watching the first election returns, and on CNN Wolf Blitzer was continuously amazed by each new vote count in Florida, exclaiming that “Trump takes the lead!”, “Now Hillary is out in front!”, when the numbers were just arbitrary depending on what precincts had reported. CNN was openly creating a fictitious back and forth foot race out of nothing, framed by ALERT graphics and dramatic music. This happened on a major network on a day of massive ratings, and the common response was, “well, they are dumb, they do this every year.” 

At the same time, over at the data journalism site Fivethirtyeight, the early returns and exit data were characterized as “excellent news for Clinton,” “bad sign for Trump,” “long night for Trump,” and so on. At the start of the primaries, they gave Trump a 2% chance of being the nominee and somehow continued to be a source of information during the general campaign, providing very detailed to-the-decimal fake precision about a Hillary lead that didn’t exist. This persisted even after returns started to come in last night, and a few hours later, the horror of Trump’s victory came to pass. And people were, to say the least, surprised.

That people (people like me: white, coastal, liberal) were surprised by what happened last night should be read as a repudiation of the media we are consuming. We’re quick to call out right wing sites as harboring misinformation, but what is clear today is that the political press, the pundits, those providing you takes, and of course all that data, down to the tenth, are also implicated in the rise of misinformation. People spent months and months clicking on Fivethirtyeight, listening to podcasts, thinking they were being informed. Super informed. It was a massive and counterproductive waste. Something we needed to come to terms with even had Clinton won is that the right doesn’t have a monopoly on political fictions presented as fact. 

…On the right, they have what Stephen Colbert called “truthiness,” which we might define as ignoring facts in the name of some larger truth. The facts of Obama’s birthplace mattered less for them than their own racist “truth” of white superiority. Perhaps we need to start articulating a left-wing version of truthiness: let’s call it “factiness.” Factiness is the taste for the feel and aesthetic of “facts,” often at the expense of missing the truth. From silly self-help-y TED talks to bad NPR-style neuroscience science updates to wrapping ourselves in the misleading scientisim of Fivethirtyeight statistics, factiness is obsessing over and covering ourselves in fact after fact while still missing bigger truths.

…An example of this came from the start of Trump’s primary campaign, when the media tried to use Trump’s outlandishness as a way to pretend the rest of it all was “truth”, that the other campaigns and their coverage were somehow in good faith. One way they did this was by calling Trump a “troll.” Trump was never a troll, he played by the silly rules of the big reality show perfectly. If we were being trolled, it was by those selling us the fiction of this election as something genuine. 

More recently, you’ll remember the tape of Trump bragging about sexually assaulting women. This was completely on-brand for Trump, but some opportunistic Republicans pretended to be just shocked by his comments so they had an excuse to jump ship from an otherwise struggling campaign. No adult learned anything new about Trump from the tape. Meanwhile, the Editor in Chief of BuzzFeed penned a victory lap for journalism, “We Told You So: The MSM, vindicated.” 

The idea that mainstream journalists uncovered facts and changed people’s minds and took a liar down was impossibly naïve…. From beginning to end, Trump was used by political journalism as an excuse to sell fiction as fact. And, in the end, Republicans came home to Trump, despite the so-called vindication of journalism; a journalism being called “mainstream” even when much of the country finds it irrelevant. 

So many more examples could be given, but it’s getting late, and one general takeaway from the 2016 Election seems clear: our popular media, from those producing it to those sorting it with editors and algorithms, are not up to the task of informing us and describing reality. This won’t happen, but those people who got Trump sooo consistently wrong from the primaries to Election Day should not have the job of informing us anymore. And if you were surprised last night, you might want to reconsider how you get information. [2]

Go read the whole thing. Jorgenson understands what is at stake. It is time to drive the lies out


[1] Michael Cieply ”Shocked by Trump, New York Times Finds Time For Soul Searching,” Deadline (10 Nov 2016).
[2] Nathan Jurgenson, “Factiness,” N. Jurgenson’s Tumblr account (12 November 2016).

Leave a Comment


If the popular vote meant anything election campaigns would be completely different, focused on getting big turnouts in a few big states. Clinton winning the popular vote is entirely due to California. Trump got more votes in the 49 other states (in total).

The polls were wrong. After the polls were proven right in the last two elections, the media was obsessed with the approach of polls + analytics = accurate predictions. There was some suspicious stuff going on with the polls, especially towards the end, but even noticing that (as Silver did) didn't change the fundamentals of the prediction. Were the members of the media reporting on the election supposed to ignore these polls in favor of some intuition that Trump was going to win? Based on some feeling that the white working class didn't like Davos Woman? A quantitative approach was always going to be the best bet for the media on both sides of the political spectrum – nobody in the media will lose status as a result of this election because they can point to the polls ("I was given bad information and I had no way of knowing it was bad"). If you had guessed Trump would win, against the overwhelming majority of polls, and he had lost, you would have no one else to blame – the blame for your mistake would be fully on your shoulders.

So the only meaningful question here is "why were the polls so wrong?"

It seems that the "shy Trump voter" phenomenon explains why the polls were wrong. That, in turn, was because Trump was demonized in the media so that people who liked his policy proposals, such as immigration restriction, lied to the pollsters about how they were going to vote. Greg Cochran predicted before the election that social desirability bias was skewing the polls by 5 points, which was probably true of several battleground states.

The national polls showing a single digit (Less Than 5) were not wrong. Most of the state-to-state polls were only wrong about Trump .
Actual, projected vote as of 30 November 2016:
MN: Clinton 46%, Trump 45% (44,765 difference)
WI: Trump 47%, Clinton 46% (24,081 difference)
MI: Trump 48%, Clinton 47% (10,704 difference)
PA: Trump 48%, Clinton 47% (64,403 difference)
NC: Trump 50%, Clinton 46% (173,740 difference)
FL: Trump 49%, Clinton 47% (112,911 difference)

What the polls said:
MN Poll Averages (Oct-Nov only): Clinton 48%, Trump 43% = 5 point difference — Both 2% off their actual.
WI: Clinton 46%, Trump 40% = 6 point difference — The Clinton % was correct.
MI: Clinton 45%, Trump 40% = 5 point difference — Clinton's avg. % was closer to her actual vs. Trump.
PA: Clinton 47%, Trump 44% = 3 point difference — Clinton avg. % was correct.
FL: Clinton 47%, Trump 46% = 1 point difference — Clinton's avg. % was closer to her actual vs. Trump.
NC: Clinton 46%, Trump 45% = 1 point difference — Trump won +4%.

Polling was essentially correct because it showed Clinton basically received the average % that was projected. Late-deciders and undecided voters went to Trump in droves, on Election Day! Or hacking?

Jorgenson has some good thinking on display but I notice in one of his examples he blithely assumes the only basis for Obama birtherism was 'white supremacy'.

That the former was stupid does not make the latter the only explanation behind it. Unless one is still in thrall to that particular assumption.