Total Pageviews

Friday, February 11, 2011

More college girl faces from the States

Here are (left-right) Alaska, Arkansas, and California. I think I'm missing an "A" state there (sorry Arizona), but here goes anyway:

Since the Latino-white-Asian populations aren't as distinct in CA as they are elsewhere, the CA composite is made from all three groups. She still comes out looking blonde, though (hey, you can't complain I'm not "representing diversity" when you're using all that hair dye, girls).


  1. Here are a few suggestions:

    Don't do states alphabetically, which would be a lot of work, do them instead in order of population and maybe points of the compass. You've done California, so do Texas, Florida, and New York next. That will get you to 30% of the population of the whole country. Then maybe do Minnesota as North Central and kind of an interesting population.

    Another way to do it is pairs of rival colleges: do UCLA v. USC, Harvard v. Yale, Duke v. North Carolina, Michigan v. Ohio St., Cal v. Stanford, Oklahoma v. Nebraska, Notre Dame v. Miami, MIT v. Cal Tech, Smith v. Mt. Holyoke, SMU v. TCU, Tulane v. Vanderbilt, etc.

    That way, you will get interest from guys who root for each college against the other, but more disinterested readers will also be able to compare across regions, across classes, and so forth.

    Another approach is to take some category that lots of girls list on Facebook or MySpace and see if composite faces look different. 19th Century novelists used to go on and on about how their characters looked. They seemed convinced that you could recognize personality traits from the shape of a person's nose. It always seemed kind of nuts to me, but you've got the tools to test things like that.

    Don't use something random like astrological sign, where any differences would probably just be luck. Try something that might somehow tie into different looks like favorite music: pop v. rock v. hip-hop v. country. I don't look at Facebook so I don't know what traits there are a big sample size of, but it would be pretty interesting if they looked different.

  2. Those are excellent ideas--I especially like the idea of comparing rival schools against one another!

  3. Try comparing girls who are "fans" of certain things (indie bands, novels, authors, movies, causes, etc.) on Facebook.

    For example, see how girls who are fans of MGMT and The Strokes (i.e., really preppy) look. Or girls who love Jane Austen. Or girls who love the movie Marie Antoinette or Lost in Translation (I think I have a few guesses there). etc.

    This way the sample size is a lot smaller and additionally controlled (because "liking" something is an pro-active act), and would be easier to do because the numbers involved are lower.