To be fair, I don't know what metrics that stat is based on. But the way I see this playing out in a more likely scenario would be like this: An above-average white woman dates 10 men; 8 of them are white because she prefers white men. Now, each of those 8 white men are above average too, because of course the woman isn't going to date down. Even so, those 8 men may have dated 6 women each, and on average only 3 of those 6 women they dated were white, and the other 3 they went for something else. So in that case, you could say that 80% of the white woman's partners were white, while only 50% of the white men's partners were white. But, 8 white men just dated a single white woman. More white men are dating white women, but the white woman is probably more likely to have a higher percentage of her partners be white.