It doesn't work. It's enormously difficult to construct a solid study with one properly isolated variable that's the subject in question. Even much, much, much easier tasks as exploring effects of smoking have been done over and over again, only to be declared invalid by yet another variable not accounted for. Statistics is hard to get right, this is why there are so many jokes about it. And trying to come up with a "proof" of anything by separating everybody into just two groups is idiotic.

Yes, it makes one feel good about oneself, and can score you some votes if you're a politician, but it's a waste of time - it doesn't prove anything and only reminds blacks and women that they're "in a group", i.e. must be different by some criteria. This is what I call "recursive racism/sexism".

It is not OK to say "first black president is a great achievement of America". The real achievement would have happened if nobody even noticed the color of his skin, you never heard of CNN anchormen talking about Bush's eye color, haven't you?