Date: Wed, 9 Feb 2000 16:49:23 -0800
Reply-To: David Cassell <cassell@MERCURY.COR.EPA.GOV>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: David Cassell <cassell@MERCURY.COR.EPA.GOV>
Organization: OAO Corp.
Subject: Re: Question about Regressions
Content-Type: text/plain; charset=us-ascii
> If I would like to regress a dependent variable Y on GDP and
> Then it should be
> proc reg;
> model y=gdp household_income;
> But what I really want is to get 2 betas for household_income. One is
> for US and other is for Japan.
> Does SAS have built in procedures OPTIONS to do this ?
You can do this with a BY statement in the proc reg.
Or create a dummy variable which is 1 for US and 0 for Japan [or the
way around if you prefer]. Then use that as an independent variable in
But *how* you use that dummy variable depends on what you want to do
your data. What is your underlying model? Your desired hypotheses?
You might have one of the following models at this point:
Analysis of Covariance
Which one you really should use will affect the outcome of your tests.
David Cassell, OAO Corp. email@example.com
Senior Computing Specialist