On Sat, 23 Mar 2002 01:45:27 +0100, "ParkStein" <[EMAIL PROTECTED]> wrote:
> Hi everybody, > > with my hope that I can get the answer here, > I send you my question. > When I'll get the SD in a data, > why do I use the square of the difference > between average and the respective value > instead of the absolute value of the difference? [ ... ] As others have said, you can use abs(diff) if you want. In addition to the good reasons that I have seen posted, here's theoretical one and a practical one: Many useful theoretical distributions, including the normal, have the first and second (central) moments as parameters. So if you want the parameters, that is what you solve for. 'Squares' offer (relatively) easy of hand-computations; which is how statistics used to be done. That is because you can decompose sums of squares, such as -- SumSquares(total) = SS(within) + SS(between). -- Rich Ulrich, [EMAIL PROTECTED] http://www.pitt.edu/~wpilib/index.html . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
