What causes people to take statistics as evidence over other
   explanations of causation? What components of statistics have
   the potential to mislead?