I don't know why I do this, but I read a lot of Republican blogs and magazines. I managed to find this piece of beauty at Culture 11's Ladyblog.

All too often on this site, we don't get to write about the non-politicall uses of data, quantitative methods and econometrics.  After the massive success of Levitt's Freakonomics, we saw a huge increase in the popularity of using economic methods to study quirky, clever or even cutesy problems.  Even non-quant people have gotten into the act.  One of the Culture 11  Ladybloggers theorized that selection bias explains why people think that jerks are more romantically successful than non-jerks.

Oh, and you know there's more.

In a nutshell, the blogger disputes the hypothesis that jerks are more romantically successful than non-jerks (I can't use the term "nice guy" without laughing, so we're stuck with non-jerk.), and says that those who beleive that to be the case are engaging in selection-bias. They're only looking at the succesful jerks, and not all the millions of jerks who are unsuccessful.  Because there are more unsuccessful jerks than there are successful jerks, she says that there must be some other variables, such as good looks, charisma, being in a band, etc., that explain why these particular jerks are successful. One of the commenters who was an econblogger decided to study the problem by means of data, and so he did.

I advise you to go read the post in its entirety, because it's a great way to see how people try to define variables and use data to solve the problem.  

DD