At one point, I was -- over-confident of the effectiveness of typing by description - especially Stratiyevskaya's.
Yet, I soon realized that
- her kind of description (raw Te data and no Ti) may not be easily appreciated by everyone
- some people do not find them useful at all to differentiate between similar types (as between SLE and SEE)
Which goes to the broader issue of type descriptions generally.
In theory, descriptions should be a "pure" form of typing since they are the immediate data-collection, real people. Yet that's also wrong. The "pure" descriptions would be individual descriptions of individual persons. No matter how long and inclusive descriptions will be, they will always go for the "average" or "lower common denominator" of a bunch of people supposedly of the same type. Yes, that "average" should be a reflection of the functional preferences and that is the point, but this is imperfect since the same functional preferences will not be manifested in precisely the same way for everyone (even if they may do so for the majority).
So even if you read all available descriptions, you will still "eliminate" some individuals who are different from the "average" of the types, which will lead such individuals not to recognize themselves as any type (they still would, of course, if they understood that makes the type be the type).
Some examples are --
- the "Fe-loaded" kind of LIIs (LIIs who are quite emotionally expressive on their own)
- the same for LSIs, where I have seen more women like that
- the "socially-awkward" kind of EIEs
- the quieter SLEs and SEEs
- the "organized" SEEs, which are often typed as ESEs by people with MBTI in mind
- the "intellectual" SLEs
- the solitary, silent, not-very Fe-active nor obviously social kind of SEIs and IEIs
- the quieter LIEs
- the "joker" SLIs (that's an interesting case, some SLIs are really into the "flickering" Fe in terms of cracking jokes of the Ne sort, but they usually don't realize when it's not called for)
- the "apparent player" SLIs who date a lot
Now how would I know that those individuals are of the types I have seen in them? You wouldn't if following descriptions. That's where a deeper understanding of what drives those individuals is important, which is obviously more difficult to spot. For those cases, you have to look beyond the usual image of the type, but with time, the true type becomes clearer.
And that's the problem with typing by comparison -- you end up automatically screening out such cases.