I think that perhaps the issue is that most people are probably taking the free online tests rather than the actual test- which I think they can be a good guidepost, and most people don't want to pay around 70-100 and something dollars understandably. But the thing is that I'm rather certain that the questions may be modeled sort of on the actual test, but most likely lean towards the typical, stereotypical view that, say, "All extroverted types must be loud!" or "All introverts hate parties!" which skews the actual definitions and recreate the definition based on tendencies.
Now, I have two friends who are ENFP that I have a very close relationship with. One of them is rather typical, I think, for an ENFP- he's loud, gregarious, is hurt if he feels you don't respect him and give him enough attention, but isn't really touchy-feely. His Ne with Fi allows him to fit in easily, socially. I don't get along with him as well, though, because his extroversion is very pronounced, and he doesn't understand why I instinctively reject social outings. However, my other friend, although still extroverted, is not quite as gregarious, and actually would prefer to sit quietly during a party and watch people's interactions to allow intuition free reign. However, we still define her as extroverted because her attention is very much directed towards other people and outside things.
The actual differences aren't surprising; we can see differences right on this forum. But, if you use the MBTI Step Two idea of facets of preferences- especially since it takes into account the the strength of the preference- you can more easily pinpoint these differences without having to question the entire type given, as it separates the tendencies from the actual definition of the preference. (And as an aside, someone criticized the test for making the E/I and J/P functions- they are actually said to be "preferences" in most explanations of Step II; "function" is still assigned to the actual Ti, Ne, etc.)
So, looking at the example I gave, the E/I facets are:
initiating/receiving
expressive/contained
gregarious/intimate
active/reflective
enthusiastic/quiet
In both cases, my friends are initiating, expressive, and enthusiastic, all in preferences tendencies. However, both lean towards a more intimate conversation style- they both speak more readily in small groups or one on one, while my second friend is probably reflective over active. However, these are simply variations- they are both highly focused on the outside world and need stimulation through social interaction, which is the actual "definition," if you will, of extroversion.
So with the sample that cheese gave: ENTP with Ti, Ne, Fe, and Se,
It could simply be that the ENTP is actually an INTP, but happens to be gregarious, or at least gregariously leaning- or lean OOP on any of the facets; the theory says that you can have up to 4 OOP facets without a preference change.
Also, another thing, I've run across the idea several times that various functions tend to strengthen as you grow older- your first function develops first, obviously, then the second around puberty/early twenties, I want to say, third around 30s, 40s, then after that the fourth function starts to develop. While your first two functions are always the strongest, it's possible, depending on development, that the second function can be rather strong, I think.
Anyways, I'm not arguing against a closer look at the MBTI- I think if nothing else, it'll be very enlightening, and of course, any improvements to make it even more accurate are great- but I think it'd help if we had some more MBTI experts to guide any improvement projects so problems can be zeroed in on, instead of trying to find problems that may already be solved in a different form.
