Thursday, April 21, 2016

Why Has The Number of NYT Headlines About Trump Fallen By a Factor of Two?

Over the past few weeks I’ve gotten the sense that, thank merciful heaven, the New York Times has stopped writing so many articles about Donald Trump. First I downloaded their data [1] to see whether this was all in my head. It wasn’t. Whether we look at headlines per week mentioning Trump (top graph) or articles per week mentioning Trump (bottom graph) we see that the NYT has been giving him much less coverage.
Why? Here are a couple ideas.

  1. Maybe the NYT, after repeatedly complaining about Trump’s free media coverage, decided to stop giving him so much free media coverage. Here’s their analysis of his free coverage on March 15 -- right around the time his coverage begins to decrease. (Nick Kristof, a NYT columnist, writes a similar piece on March 26.)
  2. Maybe this isn’t specific to Trump -- maybe the NYT’s getting bored of the primaries in general. So I looked at headlines about the other candidates. The Democratic candidates, if anything, have been getting more coverage in the last few weeks; top is Sanders, bottom is Clinton.
On the other hand, both Cruz and Kasich have also seen drops in coverage. Here are Cruz, Kasich, and Trump at the bottom for comparison.

So maybe the NYT is ignoring all the Republicans, not just Trump. (It’s worth noting that coverage of candidates may be correlated -- if you write about Trump fighting with Cruz, you’re covering both candidates.) Susan Athey, an economist who studies the internet, points out that a lot of the difference between liberal and conservative media sources is not that they cover the same topic differently, but that they cover different topics. (So if you read Fox News you’ll hear about Benghazi and if you read the NYT you’ll hear about climate change.) Given this, I’d be curious to know if you’d see a similar pattern in Fox News coverage of the Republicans.

3. Maybe the NYT is covering Trump less because he’s no longer as strong a candidate. Here are Trump’s odds of winning the Republican nomination as provided by PredictWise; Trump’s the red line. Note the fall in late March - early April.

(Or maybe causality goes in the other direction? Trump’s odds fall because he no longer gets as much free NYT coverage? Or maybe there’s a third variable -- Trump’s odds fall and he gets less NYT coverage because it becomes increasingly apparent that he isn’t worth talking about? Time series are fun.)

I don’t know which if any of these hypotheses is correct, so if you quote this post and say I’m making causal claims I will hunt you down. I’m just asking questions. Here’s a final one. No one reporter could account for the drop in Trump’s coverage -- it’s a difference of hundreds of articles. So something has to be coordinating the behavior of lots of reporters. How does that work? Is some editor at the NYT saying to the newsroom, “We SHALL NO LONGER write so many articles about the quasi-Nazi with little hands?” Or do the reporters look at what their peers are focusing on and choose to focus on that as well? Or are they all just responding to the same external factors?

Discriminating between these hypotheses is going to be hard using only data. But there are people who can help answer this question. So if you work at the NYT, feel free to explain what’s going on :)


[1] Data; my code.