Date:         Sat, 12 Apr 2008 08:14:57 +0530
Reply-To:     ajay ohri <ohri2007@GMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         ajay ohri <ohri2007@GMAIL.COM>
Subject:      Re: regression: one regressor value changes
Comments: To: Peter Flom <peterflomconsulting@mindspring.com>
In-Reply-To:  <997851.1207960274418.JavaMail.root@mswamui-swiss.atl.sa.earthlink.net>
Content-Type: text/plain; charset=ISO-8859-1

you can try regress the new variable on the old variable . or at least find
the correlation between the two.

Regards,

Ajay

On Sat, Apr 12, 2008 at 6:01 AM, Peter Flom <
peterflomconsulting@mindspring.com> wrote:

> sasworks <xuzengpj@GMAIL.COM> wrote
>
> >I have a simple linear regression with a few variables. One variable
> >has a value mostly between 1 and 200. The variable value will change
> >to be mostly between 300 and 700. Old value and new value are
> >positively correlated. The higher the old value, the higher the new
> >value. I do NOT have a sample with both old and new values. I only
> >have 2 independent samples, one has the old value and the other has
> >the new value.
> >
> >How can I "transform" the new value into the old value so that I can
> >keep the regression equation the same?
>
>
> Why will the variable change?  Is it the same variable on some new scale?
>
> Without knowing more, it's hard to say for sure, but it looks like the NEW
> variable could be the OLD one doubled and adding 300.  1x2 + 300 = 302,
> 200x2 + 300 = 700.  Of course, it could be something else entirely.
>
> HTH
>
> Peter
>
> Statistical Consultant
> www DOT peterflom DOT com
>