It definitely acknowledges our biases better than "spectrum", although I think they want varying degrees of each component, and we don't usually take a little of this star and a little more of that one.
I might've recommended "palette" if asked before spectrum got traction.
Or maybe we don't need a better word at all, maybe we should try grouping the characteristics differently such that they're no longer all together under the same heading. (I'm not qualified to say, just a hunch).
1. We've seen them long before we could really investigate them.
2. We've already grouped them up based on seeing them in a similar direction.
3. ... But it might turn out they are actually very far away from one-another, or have important differences we weren't able to see before.