Right - how do you document some subgroup without one one trying to make it into a generalizable truth about something.