It is worth examining some of that empirical evidence as to self-sorting, before going on to consider the consequences of third-party sorting or unsorting of other people. The crucial point here is that, when people spontaneously sort themselves, the results are seldom even or random, and are often quite skewed.
