Razib Khan is doing some really interesting stuff. I’d posted a link on this OkCupid info on Facebook, and now that it’s come up again, I wanted to post it here (since this repeats on facebook anyway. I’m still blown away that OkCupid is releasing (and analyzing) all this data, it’s like the fright people are having when they see some of the google search recommendations – it reveals ACTUAL biases and interest instead of the whitewashed versions people are more apt to self-report (though I understand the specific instance below is self-reporting, much of the data on the blog is based on actual interactions).
—
Posted on: October 14, 2009 5:42 AM, by Razib Khan
It’s been a few weeks, but I thought I would point you to an interesting post on OkCupid’s blog about race and dating. They report an interesting trend of preference for white men among all groups of females, as well as lots of other interesting nuggets. Of course, as many have noted the core demographic of OkCupid is a bit different than the general population, and might be especially so when it comes to racial minorities. I did think though that the stated racial preferences of the sample are interesting in light of other data which suggest women are more racialist than men.
In all classes women have stronger same-race preferences than men, except for among Asian women. In previous speed dating studies Asian women were exceptional for showing little racial solidarity with Asian men, so this follows intuition. Also, there is a healthy surplus of males on these dating sites. That might naturally modulate their stated preferences a bit in relation to women.
I am a little bemused by the OkCupid blogger’s outrage at these data. This sort of bias isn’t surprising at all from what I have heard if you’re a colored guy who swaps out your image with a white guy and leaves everything else the same in the biography and preferences, you’re in for a sweet surprise which you’ll never be able to follow up on. Next: OkCupid data shows that nerdier men get fewer responses!
