In article <[EMAIL PROTECTED]>,
Wuensch, Karl L <[EMAIL PROTECTED]> wrote:
>Well now, if the results are really normally distributed (or distributed
>otherwise but symmetrically), what is the difference between the mean and
>the median?  Nothing.

>If you are constructing confidence intervals or testing hypothesis, you may
>prefer techniques designed for means, as estimators of means are typically
>more efficient than are estimators of medians.  

This may or may not be the case.  For the case of distributions
with finite variances and finite positive densities at unique
medians, their efficiencies are comparable, and in general, one
would have to compute to decide which is more efficient.

In the case of symmetric unimodal distributions, if only one
statistic is to be provided, the median is likely to be less
sensitive to the form, as it does not depend on the tails.

If either of the conditions stated is false, the efficiencies 
are not likely to be comparable.
-- 
This address is for information only.  I do not claim that these views
are those of the Statistics Department or of Purdue University.
Herman Rubin, Deptartment of Statistics, Purdue University
[EMAIL PROTECTED]         Phone: (765)494-6054   FAX: (765)494-0558
.
.
=================================================================
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