Sure, I get your objection from an epistemological perspective; I don't disagree. The thing is that then you wouldn't be comparing "upper class women" vs the re

Sure, I get your objection from an epistemological perspective; I don't disagree. The thing is that then you wouldn't be comparing "upper class women" vs the rest of the workforce. You should be comparing upper class women in their 30s vs upper class women in their 50s vs men who smoke vs single men in their 20s vs ... In that case you might find that upper class women in their 30s are indeed more likely to leave the workforce than single men in their 20s, but less likely than men in their 50s. At that point, the "economically rational" argument becomes to only hire single men in their 20s, which I guess is what Silicon Valley does.