[Footkerchief]

The point is that the correlation probably points to an underlying variable (e.g. leisure time) instead of direct causation.