The math of beauty

[steffen]

A little treat for the weekend ...

http://gizmodo.com/5729660/the-math-of-beauty

[Ingo Mierswa]

Thanks, fun read! Would be interesting to see if the result would be the same for changed sexes - or if the behaviour patterns used for explanation in the last paragraphs might depend on the sex of the viewer.

Cheers,
Ingo

[rakirk]

It is interesting they showed game theory phenomenon may be occurring solely through the use of existing artifacts. This reminded me of the movie "A beautiful mind" (no pun intended).

[steffen]

Yes, of course this could be a case of ancedotal evidence fallacy. But nevertheless it is fun to read, and btw,  the competition-argument is reasonable (at least for me)

[rakirk]

I was thinking about this again today and I realized that they really needed to measure intraclass correlation in order to determine whether there was a construct that was initially supervised incorrectly. Classify the guys who are classifying the girls- I guess turn around if fair play

[wessel]

What makes you think they did not measure intraclass correlation?

I think these people used SPSS to do their statistics, since I recognise the scatter plot make up.
So they probably used the General Linear Model, which takes into account intraclass correlation.
Linear Regression with the beauty class recoded as four {0,1} attributes, can be thought of as ANCOVA.