Are you asking how what I described is a composite? How I'd create a composite? If I were trying to define sex, I'd try to find the most relevant attributes to human sexual dimorphism, and if and when I encountered exceptions, to me that would imply a spectrum exists and sex is non-binary.

What am I measuring? I'd be trying to find the biological measures relating to sexual dimorphism, and how they consistently they correlated to the other biological measures I listed. I'd also be looking for exceptions, which do certainly exist.

How are you translating it into a continuous variable? I wouldn't be, I imagine that upon seeing that there are exceptions to the attributes I included in my composite, where people in one or more ways fall outside of the two typical morphological/gamete/chromosomal norms, that would be what implies a continuum/spectrum. I'm not translating it, I'd just be seeing nonbinary data, I didn't translate the data to fall outside of a binary pattern or force it into a continuum, it just exists that way. Correct me if you think I'm wrong.

Why are you calling this variable sex? As I said, in my best attempt to create a composite of all traits that could be contained within a definition of sex, that's what it'd be.