[klipt]

> If men file for custody

But that's not a random sample. Given the widely perceived bias against men, one would expect only the men with the strongest cases to try for custody, and you would expect them to have better outcomes than average men.

If as an experiment you randomly selected a group of men and got them to file for custody, their success rate would probably be lower.

[lelanthran]

> Given the widely perceived bias against men, one would expect only the men with the strongest cases to try for custody, and you would expect them to have better outcomes than average men.

Actually, even on close-to-even odds, men don't try anyway (losing has a much larger effect if you're a man). See my response in this thread elsewhere for why I declined to fight when my estimated odds were 3/5.