Herman Rubin wrote:
>
> In article <[EMAIL PROTECTED]>,
> David C. Howell <[EMAIL PROTECTED]> wrote:
> On the
> >other hand, when I generate multiple independent lognormal distributions
> >and then take their sum, a P-P plot tells me that the sum is nicely
> >lognormal, not normal.
and Herman Rubin wrote:
> This leads me to suspect the pseudo-random numbers being used.
Personally, I cannot imagine any plausible flaw in a PRNG that
would lead to this result. Almost any PRNG produces "uniformly
distributed
random numbers" and the failures occur at this level - and are usually
rather
subtle. The algorithms for massaging these into "normally distributed
random numbers" are generally correct and easily verifiable.
If this happened to *me* (and stranger things have happened at sea) I
would suspect an accidentally overwritten column of numbers.
-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/ .
=================================================================
