Yes, but doesn't the genetic control here also control for that? They compared first children of refugees to their first-cousins born to non-refugees, and foun

Yes, but doesn't the genetic control here also control for that? They compared first children of refugees to their first-cousins born to non-refugees, and found a difference in frequency. So presumably a change would still be detected if sons of refugees were less likely than daughters to be hospitalized. (I can think of some other ways gender could mess this up, I guess. For instance: schizophrenia is more common in men, heavily genetic, and probably not affected by refugee status - which could create a "base load" of male hospitalizations that pushes any change among men below significance.) Also striking, though: "was not true for male offspring or the descendents of men." The headline misses it, but this result was only for daughters of mothers; even daughters of fathers (i.e. people with female hospitalization rates) weren't affected. That's the part that screams 'epigenetics' to me.