[fwip]

The NYT summary sounds accurate to me.

> > Wherever per-person Facebook use rose to one standard deviation above the national average, attacks on refugees increased by about 50 percent.

> That sounds horrible, but it is actually a claim about variation across municipalities, not a claim about the absolute importance of the internet.

That's why they used the word "wherever," to emphasis that it was based on locality.

That seems to form the bulk of Mr Cowen's objection. The remainder appears to be recitation of the stock phrase "correlation is not causation," mock horror and snark about the methods the study used (Nutella! heaven forbid!), and expressing his general confusion about parts of the study he doesn't understand.

To my eyes, this doesn't appear to be a "comically erroneous study," unless, as Mr. Cowen does, you find great intrinsic humor in the concept of Nutella. Nor does the NYT reporting on it seem to misrepresent the study any more than articles for the layperson usually do.