This isn't the first time I've wondered about correlation of surveys. One of my crazy if-only-I-could-retire-now ideas has been a site like Emode or Quizilla, with all results data (anonymously) evaluated for reliability and correlation with each other and demographics.
Since I first dreamt that up, though, the blog phenomenon has surfaced, along with hundreds of people taking all sorts of self-administered tests and surveys and posting their answers or measure results. Granted, many of these are meaningless "which character are you" surveys, but even then the raw answers might be scientifically interesting.
What I wish for now, then, is a more structured way to work with all this data. Perhaps if more measures were tracked by FOAF as MBTI is now, so a client could mine for correlations. If only I could get the blogosphere to move off social network islands and claim ownership of their own information