Let's put it this way: you've put people into buckets. Great. Science says: if these buckets represent something inherent about the person, then you should be able to predict using this information.

Maybe you can predict the likelihood of someone getting depression? Or of having a high paying job? And those predictions have to be better than random guesses, and hopefully better than existing tests for those things.

So, where is the scientific evidence for Myers-Brigs? What exactly does the typing predict about a person that we can't predict better using other methods?