Date:         Sat, 14 Nov 2009 06:45:45 -0500
Reply-To:     Peter Flom <peterflomconsulting@mindspring.com>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Peter Flom <peterflomconsulting@MINDSPRING.COM>
Subject:      Re: Sample of a bivariate Probability Density Estimation data
Comments: To: Stephane COLAS <Scolas@DATAMETRIC.FR>
Content-Type: text/plain; charset=UTF-8

Oh I see.

In that case, you can generate numbers from a lognormal; that's a parametric distribution.

So, find the key parameters for your population (I believe it's just mean and sd and (since it's bivariate) the correlation) and then generate numbers from a similar distribution.

I don't recall exactly how to do this (I usually do this sort of thing in R, not SAS) but it's not hard.  I am sure it's in the archives somewhere.

HTH

Peter


-----Original Message-----
>From: Stephane COLAS <Scolas@DATAMETRIC.FR>
>Sent: Nov 14, 2009 4:38 AM
>To: SAS-L@LISTSERV.UGA.EDU
>Subject: Re: Sample of a bivariate Probability Density Estimation data
>
>I want thé third proposition : the sample distribution must follow the population distribution. Ie.
>If the population follow a lognormal distribution, I expect that my sample shows the same.
>If i use surveyselect, the distribution is not followed. So ?


Peter L. Flom, PhD
Statistical Consultant
Website: www DOT peterflomconsulting DOT com
Writing; http://www.associatedcontent.com/user/582880/peter_flom.html
Twitter:   @peterflom