[jprete]

Nitpick: The five-factor scale (extroversion, agreeableness, openness, conscientiousness, and neuroticism) both reproduces and makes good life predictions.

MBTI and most other personality tests unfortunately seem to be astrology for the scientifically oriented.

[nostrademons]

The correlation between MBTI and the Big-5 is surprisingly strong (except for neuroticism, which has no representation in the MBTI):

https://en.wikipedia.org/wiki/Myers%E2%80%93Briggs_Type_Indi...

I look at personality typologies in general as approximations. If you actually want to get to know a person, get to know them. But when you need to make a snap judgment about how someone might see the world and react to you, on very little information, it's handy to have a general archetype as guidelines. You can fill in all the details later.

It's not unlike Carmack's fast inverse square root, a Bloom filter, or how Google hasn't used PageRank since 2006 (instead substituting a cheaper-to-calculate approximation). Yes, they give wrong answers. But the answers are usually close enough that when you lack the computing power to get a better result, they'll do.

[HelloMcFly]

The biggest issue w/ MBTI and others like it is the desire to put you into a category as an outcome. This is really bad for reliability of the test - take it again and change just a few answers, you may be in a different category altogether! But it's often great for sparking team discussions on norms, behaviors, etc. But the sheen of science and validity can be misleading.