"Robert J. MacG. Dawson" wrote: > > There are several different definitions of skewness, which are not > mathematically equivalent. This is clearly not a Good Thing;
I claim it is an essential thing; just as there's no single best measure of location, no single best measure of spread, no single best measure of monotonic correlation, there's no single best measure of skewness. So we need more than one measure. The big problem is when people call a particular measure "skewness" instead of "a measure of skewness". Glen . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
