Do they separate causation from correlation, or is that not the point?