""You mean the theoretical *increase* would be *four hundred* per cent, because then *all* the fatal road traffic accidents would be caused by drivers who've been drinking.""

Sorry Lighter, but I mean nothing of the sort.

My conclusion is valid (though in a mad sort of way), since I stated that the absolute total of road deaths would decrease by eighty percent. Correct?

You drew the conclusion about the proportion of those deaths which would be due to drunks all on your own, and at one hundred percent (give or take a couple) you were correct in that.

This is the lunacy of statistics, where two different viewpoints give seemingly incompatible results simply because neither understands what the other said.

Politicians are very good at misusing statistics in this way, though their "misunderstanding" is often deliberate.

As somebody once said, "Tell me what you want to prove and........"

Don T.





""The sample size matters enormously, as do its randomness and representativeness.""