i don't know if this information will be helpful but most personality quiz type inventories, even those used in Real Scientific Research (TM) are limited in their applicability. All defined personality "types" can be overwritten by the context of the situation you're in and are, at minimum, always influenced by context. This is why the "better" inventories of this type actively tell you to limit the context of your responses to be only when you're at work. Plus, self-report data is... easily biased as you have insight into your intentions in communication whereas an outside party doesn't have a clue what you intend.

Which is to say: your conflicting results aren't surprising. That happens a lot with this stuff. I would suggest keeping the resultant advice that seems useful and toss out the rest.