Report on Scourge of Fake News Turns Out to Be Faked

I’m sorry, but this “fake news” thing is too funny and so so typical! From Paul Crookston at National Review:

Buzzfeed is coming under fire for the methodology of its Facebook exposé.

The presidential election proved to be fertile ground for the growth of fake news stories. As people became annoyed by tall tales such as “Pope Francis Endorses Donald Trump for President” popping up in their Facebook feeds, the mainstream media decided it was time to fight back. As calls for action from Facebook grew louder, Buzzfeed released a bombshell report: Fake news outperformed real news on Facebook in the final months of the election. This was then widely taken as fact, but it turned out to be, well, untrue.

As Tim Carney reports at the Washington Examiner, Buzzfeed’s research methodology was a mess:

The BuzzFeed study looked only at the top 20 election stories, in terms of engagement, by fake websites and compared it to the top 20 election stories from a tiny list of mainstream sites. The list included left-leaning opinion-heavy outlets like Huffington Post and Vox. . . .

So the “Real News” numbers are from an incomplete, odd, and unexplained subsample of the media.

Buzzfeed narrowly defined “real news” to exclude widely read sources such as Yahoo News, Reuters, the Daily Mail, Associated Press, and all newspapers outside New York, Los Angeles, and Washington, D.C. This skewed sample undermines Buzzfeed’s claim that fake news “outperformed” real news on Facebook, since it hides the total amount of traffic that real news sources received, furthering the idea that mass exposure to false stories led to Donald Trump’s election.

Since Buzzfeed published the story, its author Craig Silverman has acknowledged that the data are not perfect, and he stated on Twitter that no one should conclude fake news won the election for Trump. But he stands by the conclusion that fake news saw “bigger engagement” than real news during the campaign. This narrow definition of “engagement” is misleading, to say the least.

Read more: National Review

Image credit: Mohd KhairilX / Shutterstock.