Well, the whole 3rd paragraph of the article is talking about how a poorly implemented culture fit criterion can be a place for bias to hide. I would assume tha

Well, the whole 3rd paragraph of the article is talking about how a poorly implemented culture fit criterion can be a place for bias to hide. I would assume that that was meant to mean all kinds of bias, not just bias that has nothing to do with any protected classes. The survey itself was all qualitative, self-reported data. People aren't typically aware of their own racial, gender, etc. biases, so one wouldn't expect to see that being something they up and volunteer in their responses even if it is a factor in how their "culture fit" screen works out in practice.