For what it's worth, I strongly recommend separating your investigation of functions-theoretic typologies and ones that are more empiricism-oriented (like the Big 5).
A good example of why this might be is that, while you *could* argue the attitudes of dom and aux should be reversed to provide balance between the subjective and objective factors in cognition -- although orthodox Jungian typology has them in the same attitude, eg Nietzsche Chapter 3 Psychological Types is shown to be Ni+Ti, it seems like the I/E dimension that corresponds to what's measured by the Big 5 is hardly this way (you don't need much balance between the I and E sides).
But more to the point, they're simply different systems on closer inspection. So, assuming you're deciding say, between NiFe and FeNi, I'd just make that a pretty independent consideration....e.g. I know I would test as an introvert in normal tests, probably something like INTP.
With regards to functions tests: when tested in that format, one really can't expect the orders of the function-attitudes to be anything particular; in my experience (or by the data -- the Jungian analysts Singer and Loomis actually built a test, and it confirms this), to get at one's axis eg Fe-Ti or Ne-Si, one actually has to construct in one's mind a new variable called, say, Ne-Si (rather than treating Ne and Si as two separate variables.).
Part of the reason for this is that the very idea that, say, someone is Ne-dom/Si-inferior is a philosophical one: Si-inferior does not mean you "use Si and don't use Se" in any straightforward way -- it's more like Si is relegated to the unconscious/compensatory to Ne (so you'd detect it indirectly in the philosophy of ego-consciousness always ending up played up against an Si attitude to define itself against), so somehow, there's no real sense you can empirically test for "using Si" independent of Ne to establish Ne-dom/Si-inf -- you have to examine Ne-Si as a variable of its own right. Same for Fe-Ti, Fi-Te, etc.
Basically, make a function axis a variable rather than making the function the variable.
I can't see your post history, or I'd go find previous type-me's to see if there's anything that points me one direction, so I'm just rambling these points of theoretical concern, as I've found they're at the heart of what is unclear about the systems.