The study was posted on FR a couple of years ago.
I didn't keep a link to it. But if you're interested enough, you could probably Google something up. Suggest "military (or Army) demographics".
Well, I must say, I am disappointed in you. You tell me that I’m dead wrong and yet you don’t bring in the evidence.
So, I guess I’ll stick with what I thought in the first place. Intuitively, it seems perfectly reasonable.