Skip to content

“Social Media and Fake News in the 2016 Election”

Gur Huberman asks what I think about this paper, “Social Media and Fake News in the 2016 Election,” by Hunt Allcott and Matthew Gentzkow.

I haven’t looked at in detail my quick thought is that they’re a bit too “mechanistic” as the effect of fake news is not just the belief in each individual story but also the discrediting of the whole idea of independent sources of news.

Gur adds: “Plus they are estimating a weak effect on (probably) a small subset off the voters who tipped the election.”

The above is not intended as a “debunking,” just some concern about how these very specific findings can be overinterpreted.

One Comment

  1. Terry says:

    The paper by Allcott et al. appears to be rather results-oriented. The authors identify a subset of the dishonest information voters received that was, on average, pro-Trump and they analyze the effects of that subset. The results, therefore, tend to be anti-Trump. The authors could have looked at all the informational dishonesty and bias during the election, but they chose not to. They do not even mention how the results of a larger analysis might be different.

    Of course, the big-name media were predominantly biased against Trump, and so the paper does not address the much more interesting question of how the totality of all biased and dishonest information affected the election. To put it another way, what would the election results have been if media coverage had been “fair”. (A very tall order, I grant, and fraught with many difficulties.)

Leave a Reply