It seems that, in the end, there's so much variation that typing by observation is precarious at best.
I personally think that Kiersey does an awesome job of connecting type-related cognition to behavior and other observable phenomena, but that's another story.
Basically, typology enables you to "take a peak" into the minds of other people. But in a way, you kind of have to take it on faith that the data is reasonably accurate.
Unless a shitload of people here are mistyping themselves (which I consider a possiblity) I think this forum proves that people who have very similar ways of perceiving and judging can still end up being very different people.
It's sort of like that quote in Jaguar's siggy:
“Learn your theories well, but put them aside when you touch the miracle of a living person” --Carl Jung