Problems with that paper. If OKCupid addressed these in their data set, the didn't say. 1. OKCupid is a worldwide site but they don't say whether they're corre

Problems with that paper. If OKCupid addressed these in their data set, the didn't say. 1. OKCupid is a worldwide site but they don't say whether they're correcting for that by limiting to the USA 2. They don't say anything about correcting for regional differences. 3. They say nothing about the message rates. What percentage of a black woman's incoming mail is from white men? If it's some small percentage, can one discount the novelty effect? 4. They don't correct for income. Since in the US, black men have the lowest income of any group, it's much more likely that a given the choice between a black man and a white man, the white man will have the higher income. 5. Likewise for education level of the sender. 6. Likewise for message content. Is it possible that black men send messages with different content than white men? 7. Finally, this only considers message rates, as if the goal of a dating site is only to get other people to return your messages. This isn't any kind of a scientific study. Please don't act as though OKCupid has proven anything significant.