A Different Game

To receive a free full-text email of The Zeitgeist whenever we publish to the website, please sign up here. You’ll get two or three of these emails every week, and your email will not be shared with anyone. Ever.

The Zeitgeist on the day after a presidential inauguration is unlikely to catch anyone by surprise. It was the only topic of any significance in political news. It was the only topic of any significance in market news. The language used to describe this inauguration was central to nearly every topic over the last three days. Even sports.

The quantity of that coverage which would fit squarely into what we call Fiat News – news which is designed not to inform but to instruct the reader how to think about an event or topic – would also be unsurprising, I think. We could not shock you with an argument that the editorial side of nearly every national news outlet has been drawn into one end or another of our polarized politics. You would not gasp, I suspect, to learn that this often bleeds into the newsroom.

So I don’t expect you to do more than shrug that this was the inauguration coverage angle chosen by the New York Times, promoted on social media with the following, um, austere copy:

“Whether or not related to the former president’s absence, a bipartisan lightness seemed to prevail across the stage at President Joe Biden’s inauguration. Snow flurries gave way to sun.”

Washington Breathes an Uneasy Sigh of Relief [New York Times]

The piece is part of a feature series on the inauguration, but it is clearly being positioned as news, or at least news-adjacent. It’s got Washington in the dateline. It’s not marked as an editorial or opinion content in any way. And yet it contains descriptions of events which “evoked a mood”, which were “the most striking”, which “felt like a friendly gathering”, which “conjured a sense of respite”, and of “lightness which seemed to prevail across the stage.” In a helpful definition of Fiat News, the author offers a literal translation he’d like you to use for Biden’s speech: “Phew.” As it happens, I happen to agree with the author, yet I enjoy the peculiar benefit of realizing that this is an opinion and not a fact.

The usual suspects on the other side of the political spectrum are in the same business, of course, just in a smaller form factor. This brief Fox News article manages to jam in a lot Fiat News language. In only a couple hundred words, the reader gets multiple scare quotes, an “appears to be”, a subjectless “perceived” and multiple references to the loaded expression “pushing a progressive agenda.”

So if you were hoping that Trump’s departure would herald the triumphant return of the national media to its historical role as an agent for the people rather than a principal acting on its own account, your watch is not yet ended. Sorry.

All the same, I’d like to pose a brief question. If you were asked, “Was national media coverage of the inauguration of President Joe Biden more or less positive and friendly than it was for the 2017 inauguration of President Donald Trump,” what would you say?

Again, I think I can hazard a guess.

But by one objective – and woefully incomplete – lens, they were almost exactly the same.

I am talking about the lens of calculated sentiment, by far the most common method by which those commenting on markets, scraping news and scraping social media evaluate the nature of opinion language.

Using the standard sentiment scoring library from our friends at Quid, we estimate that the mean sentiment of 2017 coverage of the Donald Trump inauguration by national US media outlets from the day before to the day after the event itself was actually 15% MORE positive than the 2021 coverage of the Joe Biden inauguration.

What are the possible reasons this could be the case?

1 – Maybe media actually were more positive in 2017 coverage than they were in 2021;

2 – Maybe the sentiment library isn’t well-suited for this kind of analysis;

3 – Maybe the coloration from the connected Trump-related and COVID-related news made 2021’s coverage more negative than it would otherwise have been;

4 – Maybe affect creeps in not so much through explicitly positive or negative language, but primarily through framing, editorial coverage decisions, non-affected language of meaning and the decisions to include or exclude certain information from a news article;

5 – Maybe most news outlets and publishers have long since realized that the metagame – the strategy which survives repeated play – that permits shaping how people think about the news without them being acutely aware of it requires outlets to shy away from explicitly “positive” or “negative” biases in their language in favor of these techniques.

I won’t tell you what to think, but I’ll tell you what I think. I absolutely don’t think #1 is the case. That flurries-gave-way-to-sun nonsense wasn’t a New York Times exclusive, y’all. I think #2 and #3 probably are true – at least a little bit. But #4 and #5? I think they are the real story here. That story isn’t new. And I think that story has two lessons for us:

First, don’t expect the creep of editorial bias into news to hit you over the head with big, opinionated-sounding language like the articles referenced here. Framing and the willful or subconscious inclusion/exclusion of facts are the tools with which news is shaped into explanatory news without looking like it.

Second, if someone is selling you on natural language processing-based research about markets, politics or anything else that is built on the shifting sands of a calculated sentiment and nothing else, it is unlikely that what you are getting is useful.

A game in which the players know the rules against which they have been measured is a different game.