I've seen many people taking the MBTI function orders and such as a pattern and then trying to rationalise the pattern against itself, looking for areas where the pattern does not match and so on. Now is it just me or is this a mistake? Surely the pattern is resulting from the study of people. The pattern itself is a result and not the primary source, ergo arguing over where it follows a trend and where it does not can only be done out of interest voiding any revelations stumbled upon during the investigation.
I only highlight this as I've had many discussions where I get the feeling that people are basing their assumptions and advice more on an investigation and understanding of a pattern of results than on actual observation of the primary material, ie people.
I'm not saying it's wrong necessarily, just challenging the approach.