In England during WWII, Air Force chiefs went and examined some of the British war planes that had been hit by shrapnel to determine which areas were hit most often, so that they could put armor plating on those areas. A statistician spoke up and said “Wrong”.
Why?
Oh goody. Dickhead riddles again.
Russell
Simple: the planes they were observing were by definition NOT the ones that were destroyed in flight or crashed on bombing raids. The planes they were looking at had made it back in one piece, so a study of the injuries to those planes would tell you where NOT to put armor – a list of the non-vital spots on a plane, so to speak.
If they wanted to find out where to put armor to prevent crashes and fatalities, they’d have to look at planes that DID crash, most of which would unfortunately be located in enemy territory and/or blown to smithereens.
Classic example of why self-selected samples are often the worst for study.