From my understanding, this Science paper does not compare the spread of real news vs fake news, but rather of fake rumors vs real ones - this is an artifact of their data collection methodology theatlantic.com/technology/arc… 1/
In essence, they extracted Twitter threads where someone posted a link to a fact-checking website to either confirm or debunk the rumor. Such event are fairly rare, since they only discovered 160k theads over the entire lifetime of Twitter (10+ years). 2/ 
Still, the finding that fake rumors spread quicker/more than real one is interesting, but this has nothing to do with news published in authoritative news outlets 3/
While it is reasonable to assume that ppl are more likely to share dodgy news than authoritative ones, this study doesn't provide evidence in that direction - more work is needed - good to see that there is a growing community (from many disciplines) studying the problem 4/4
Link in first tweet seems broken, here is the working one: theatlantic.com/technology/arc…
