I took yet another one of those Jungian personality type tests today.

[randomness]

(They always remind me of wolfkitn, actually, because she was the first to adminster one to me, at least in person.)

Anyway, I'm a ENFP. I'm not sure anyone who knows me reasonably well is at all surprised by this.

Comments

[yaleartificer.livejournal.com]

There's a section in "The Big Test," by Nicholas Lehmann, about the history of the Myers-Briggs. Basically the guy who started ETS was so pumped about the success of the SAT that he wanted ETS to have a test for personality, too. So he fell in with this failed novelist woman that provided him with just such a test. Only problem? No basis in science whatsofreakingever. Over the years ETS kept assigning psychometricians to the test to work on it, only to have them throw up their hands in disgust at the thing.

One tipoff is that the test acts as if these personality traits are boolean with this huge cutoff in the middle, when they're probably somewhat normally distributed. This is why so many people end up with X's for personality traits ... and yet, all the writeups for the various personality combinations act as if this is a rare case instead of the norm.

Even the best personality tests tend to be iffy, though, because people act differently under different circumstances, and because there are mathematical results that suggest that the traits you choose to put in your personality test can be pretty arbitrary, and you'll still get about the same amount of clustering and repeatability.

[r-ness.livejournal.com]

Over the years ETS kept assigning psychometricians to the test to work on it, only to have them throw up their hands in disgust at the thing.

Cute! I do think they've become so much a part of pop culture that they're a phenomenon of their own.

Even the best personality tests tend to be iffy, though, because people act differently under different circumstances, and because there are mathematical results that suggest that the traits you choose to put in your personality test can be pretty arbitrary, and you'll still get about the same amount of clustering and repeatability.

I had suspected something like the latter--well, the former, also--and would definitely be interested in any links if you have them handy.

(It would be amusing to put together a test with arbitrary traits and then post it as a meme...)