[EMAIL PROTECTED] wrote:
>
> > What would be good similar examples for
> > distributions that are symmetric but not normal
Tosses of a "fair coin", of course.
If we mean "approximately symmetric":
I would guess that heights in an adult population would be a reasonable
example of a short-tailed distribution (mixed, two means, one SD). Here
we have essentially a convolution of the two "canonical" symmetric
distributions.
I seem to recall the titration data in one of the MINITAB data sets
(ACID? TITRATE? Some name like that) had a heavy-tailed, roughly
symmetric distribution - presumably a mixed model with one mean and two
SD's. [those who did it correctly and those who messed up].
For a smilar model: final position of a golf ball after a putt.
I think the Old Faithful time series is bimodal with one hump bigger
than the other, but selecting eruptions based on appropriate previous
history may give a reasonably symmetric two-hump conditional
distribution.
-Robert Dawson
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
. http://jse.stat.ncsu.edu/ .
=================================================================
