If (say) 15% of drivers are speeding at any one instant then the drivers who DO speed are LESS likely to have a fatal accident and MUCH less likely to have an accident at all.
Likewise IF 10% of drivers at any given instant were over the legal limit and only 8% were involved in fatal accidents you would HAVE to conclude that drink drivers were safer than sober ones. BUT, in fact the number of drivers over the legal limit at any one instant is MUCH less than 8% so you have to conclude that drunken drivers are less (much less) safe than sober ones.
The logic is exactly the same - but the figures are different and the conclusion is therefore inevitably different.
The fact there are more speeding motorists than drunks is a TOTAL irrelevance - it's the comparison between the number driving illegally and the number who have accidents that is the ONLY matter of consequence in judging how safe they are.
Is that really such a complicated thing to understand? It's not rocket science, its the most basic of statistics.
|
|