That is assuming everyone who is one of those groups first sees themselves as marginalized. imo it would've been better to not have that first question and just ask people which group they identify with.

That’s a fair point, though I would be surprised if it made a huge difference. The actual survey question was: “Do you consider yourself a member of a group that is underrepresented or marginalized in technology?” I would expect most members of a minority to consider themselves underrepresented just by virtue of being a minority. And as the choice of groups was explicitly given, I assume that the survey allowed to view the groups and then go back to answering the first question.

I'd expect that as well, but we can't really assume that.

Being a minority does not make you underrepresented. Underrepresented means there are fewer than you'd expect, given population-level numbers. In the Rust community it certainly seems true that trans people are overrepresented, though "marginalized" almost certainly still applies. The same goes for LGB, which again does not seem underrepresented in the tech community compared to society writ large, and I think many LGB people probably don't see themselves as "marginalized" in 2025, but I could be wrong.

> I would expect most members of a minority to consider themselves underrepresented just by virtue of being a minority.

I don't see why that would be the case.

And for tech in particular I'd say women (half the population) are underrepresented and LGBT (a definite minority) are not. Marginalization is a bit more complicated but similar.