Okay, I'll take a shot at it. The man is more likely to have two sons.
I heard somewhere that a given man produces male/female
children with unequal probabilities, so that if the man has one male
child another is (slightly) more likely to be male than female. If we
knew that the woman's two children hade the same father, the probability
that her other child is male would be the same. But since her two
children could have different fathers, that would bring her probability
closer to 50% so the man wins.
I have two questions:
(1) Does it matter whether it's the older child whose sex we know? Or is
this a red herring?
(2) How is this related to nuclear power plants?
--
Martha Takats
Department of Physics and Astronomy
Ursinus College
Collegeville, PA 19426