Jakob Nielsen's Alertbox, March 6, 2006: Outliers and Luck in User Performance
Almost all statistical analyses assume that data follows a normal distribution (the famous bell curve). Most people take this on faith, because it's true for so many phenomena. But let's check.
Yes, let's.
Addendum 11/4/06: Claus Cyrny sent me this interesting link -- homelessness follows a power-law distribution, not a normal distribution. Well of course it does. The article makes a similar statement about car pollution. The point being that dealing with the tail end of the homelessness curve or the pollution curve will have a disproportionate effect on the overall problems.