Back to questions above about the nature of facets, behavior versus preferences, etc., the facets were meant to cull out different aspects of the dichotomies. Myers was working on a way, at the time of her death, to talk about the differences we see within whole types. The way Step II was developed, the facet midpoints, etc., were compared only among the representative sample on each dichotomy. So all E's and I's scores, no matter their whole type, went into determining the norms for each facet. But, your report shows how your results compare to others of your own type. It's still a self-reporting instrument, not a diagnostic tool, so again people can report as they should or want to behave. That said, in my experience people understand their results. They almost always agree with "high" scores and easily explain midpoints and out-of-pattern results (an E who has an I facet score on one of the facets...etc.)
They are meant to explain underlying aspects of the dichotomies. For example, Extraversion and Introversion are fundamentally about source of energy--from action and interaction or from reflection. But we all know that they're expressed in different ways. The facets try to get at that. And they're very specific.
Initiating-receiving is about our basic orientation toward communicating and connecting with others.
Expressive-contained is about communicating emotions, feelings, interests and experiences
Gregarious-intimate is about the breadth and depth of one's connections to others
Active-reflective is about how you engage with your general environment for fun, socializing and learning
Enthusiastic-quiet is about the level of energy one brings to exchanges with others
...and there are other facets of E-I, no doubt, that aren't covered by these.
All the facets are this specific in nature. It's impossible to use the instrument well without deep study. And it has, naturally, the same limitations as any self-reporting instrument. That said, it is useful when used in the right conditions by people who understand its purpose, limitations and correct interpretation.
I'd have to direct further questions to the manual. As I said earlier, I use it occasionally, I train others to use it. But...always, always, I'm far less interested in instrumentation than in application of the theory. The instruments are only tools to introduce people to the theory so we can get down to the hard work of making constructive use of differences. In some circumstances and with some populations, the instruments do not work well (and the manuals detail much of this), but the theory can still be applied...