Context is everything: I work with data and methods and science, and I get paid with grant money. My motivations in publishing openly are to ensure other researchers and data enthusiasts can read and use my stuff, so my concerns are very different from, say, an artist’s concerns. If someone does something cool with my data and cites me, that’s a good thing. If someone takes an artist’s awesome image and sells it on journal covers without telling anyone, that’s not okay. My answer comes from a data/science perspective.
Most of the vocal open access people in my field go for CC-BY, though I’m curious about why you’d object to the noncommercial clause. I can absolutely understand why an academic might prefer that…and I can imagine an institution requiring it (because if it can be monetized, there’s a whole tech transfer department in most unis that will want a piece of it first).
I have a personal preference for SA, but I’ve noticed there’s not a lot of love for it.
Because reputation is important in research, my answer is kind of based on the zeitgeist in the field. I’d tend towards the least restrictive license I can get away with given institutional and publication guidelines, and it looks like it’s possible in many cases to go straight for CC-BY.