crashfrog writes:
My other favorite misuse of stats is drunken driving. If 30% of accidents are caused by drunken drivers, aren't sober drivers more dangerous since they make up the other 70%?
Only if there are a comparable number of drunk drivers to sober drivers. The statistic is a bit meaningless as it is, but I would imagine that sober drivers outweigh drunken drivers by more than 3:1. If this is the case, then you have a small number of drunk drivers causing a disproportionate number of the accidents.
™ Version 4.2
