Date: Tue, 4 Oct 2005 09:26:24 -0400
Reply-To: Venita DePuy <depuy001@NOTES.DUKE.EDU>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Venita DePuy <depuy001@NOTES.DUKE.EDU>
Subject: Re: draw a normal distribution (also lognormal)
Content-Type: text/plain; charset="US-ASCII"
You can also use the method Rene mentioned, in Proc Univariate instead of
do i = 1 to 100;
j = rannor(i);
k = exp(j);
symbol v=circle cv=blue;
*to test distribution;
proc univariate data=file1;
var j k;
QQPLOT / normal (mu=est sigma=est);
one more question, how can i use QQ plot to check the mormality of
sample in this case?
On 10/4/05, Venita DePuy <firstname.lastname@example.org> wrote:
> I believe the following will create a normal distribution (with mean 0
> std dev 1) Note that it looks much more normal if you use 1000 points,
> If you want to get particular, you can have Proc Univariate draw in
> for what a normal and/or lognormal distribution, with the mean/std dev
> the data, should look like.
> Hope this helps.
> j is normally distributed.
> k is lognormally distributed.
> You can do a simple transformation for a lognormal distribution.