And? How do you get from the existence of those differences to observed differences in representation? That's a huge conceptual leap.