> The problem is that if you're going to make an inflammatory argument It's not really inflammatory. The statistics say black women are generally more obese. I

> The problem is that if you're going to make an inflammatory argument It's not really inflammatory. The statistics say black women are generally more obese. It's just another suggestion for the difference in responses (which has been established as being wrong for the statistics OKCupid are using - it could still be generally accurate..) > implying that OKCupid is censoring the real reason for the apparent bias against black women on the site: they're just fat. Nah.. I can't see any suggestion that OKCupid was censoring information? I think he means the blog post this links too. Also: there is not implication that "they're just fat". Apparently statistically Black women are more obese than White women (I assume those are US stats); it's a reasonable suggestion that the OKCupid community mirrors wider society. > it claimed that a pairwise comparison between black women and white women could explain black women's poor response rate as compared to all races. That doesn't make any sense. Well we are mostly comparing this to the response rates of white women (at least that's how I am reading all of this) so I see no problem in the conclusion. White women have the highest response rate so it makes logical sense to use them as the comparison point. If we want to consider other races, such as Hispanic, then, yes, we would need the data for them too (who knows; it could match or it could mismatch - therein is the beauty of statistics).