[jdmichal]

This is classic correlation is not causation. The thing about correlation is that it could be a causative relationship, or there could be another set of untracked variables that's causing some or all the effects, or it could be unrelated coincidence.

Now, maybe this is a difference between the study and the article. Maybe the study makes stronger claims here than the article does. But I didn't see anything in the article that claimed nor demonstrated causation, only correlation.