I think there is some room in between those two.
At a minimum you'd expect a few more companies, more sources than just code review and code productivity metrics (this alone disqualifies the study because it centers on just one task: software development) etc.
I have not made up my mind either way so I'm not sure where you pulled that from. I even wrote in another comment upthread: "It is entirely possible that these conclusions (which by themselves are not all that shocking or novel) hold true over larger samples and across multiple types of company but that's not what they did. "
But you probably missed that.
Would you be saying that if you agreed with the findings?