Type statistics is a common reoccurrence in MBTI, and the rarity of certain types often comes up in discussions. However, throwing around numbers, statistics and how many of a certain type you may or may not know, each has their own problems.
First of all, many of the statistics out there are based on tests. I doubt this is news to most people – but tests are bad. Even the official one, Myers Briggs and Keirsey temperament sorter, are based on dichotomies, something that for starters doesn’t accurately correlate to the cognitive functions. Additionally, in MBTI, it’s based around preference rather than the strength of the functions, and in Keirsey temperament sorter, it’s based around his four temperaments. It’s a bad base to rely on when it comes to statistics.
The other approach to seeing how common a type is, is to look around and see how many of each type that we meet. But for this to work, everyone around us has to be accurately typed, which is unlikely in itself. There’s also a danger of confirmation bias, where we type people as certain types based on a subjective preexisting idea of how common a type should be. The second problem with this method is that we have no idea whether the numbers we get are representative for the entire population, whether in a country or the world. There is a selective bias of people we get along with, the subculture we are in, etc.
Lies, damned lies, and statistics. If we know how common certain types are, what would it help? To use it in deciding the type of someone, or even of a group of people, is a simple fallacy. Our group of friends or the context we’re in might, as mentioned before, draw a certain crowd.