there is a measure usually called AD ... average deviation ... this is the 
average of the absolute deviations around the mean
perfectly good general indicator of spread
problem is ... can't seem to be used for anything else ... BUT, if all you 
want to get is a measure that will compare several distributions on 
variability ... it will generally do fine

the variance on the other hand ... is like one step up the chain ... 
average of the SQUARED deviations ... but, one problem with it is ... hard 
to interpret the actual calculated value ... ie, if the variance of test 
scores is 42 ... this means 42 SQUARED TEST SCORE UNITS ... most folks 
don't know what a squared test score means (nor squared pounds, etc.)

taking the square ROOT of the variance essentially gets the variance back 
to units of the original scale ... and in this form ... sqrt of the 
variance, it is called the standard deviation

At 01:45 AM 3/23/02 +0100, ParkStein 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?
>
>why root (sigma (average - x(i))^2) ?
>why not sigma |average - x(i)| ?
>
>thanks
>
>
>
>
>
>.
>.
>=================================================================
>Instructions for joining and leaving this list, remarks about the
>problem of INAPPROPRIATE MESSAGES, and archives are available at:
>.                  http://jse.stat.ncsu.edu/                    .
>=================================================================

.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to