View Single Post
  #46  
Old July 6th 03, 07:19 AM
Craig Prouse
external usenet poster
 
Posts: n/a
Default

Statistics can still be useful even though individual
variation makes them less than perfect predictors of risk. Plus, people
might as well assume they're no better than an average pilot, because half
the people will be right.


Statistics are of course most useful when they are abused. One common
technique is to deliberately confuse the difference between the mean and the
median of a data set.