Even if those correlations hold true and have the effect you're describing (I'm far from convinced that's the case, but it's moot), it strains credulity by a mile to claim that this is some byzantine plot for "backdoor age/sex/income-discrimination". That would mean that those making decisions about open offices are 1) thinking through the complex and frankly tenuous connections that you're describing, 2) finding them credible, and 3) deciding that the collateral damage to other employees' productivity and happiness is somehow worth the weak effect this would have on furthering discrimination. There are an infinite number of ways to backdoor-discriminate on those bases that are not very noticeable, far more effective,