Date:         Wed, 30 Nov 2005 11:13:30 -0500
Reply-To:     Wensui Liu <liuwensui@GMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Wensui Liu <liuwensui@GMAIL.COM>
Subject:      Re: multiregression question
Comments: To: lixiaoyao <lixiaoyao5880@yahoo.com>
In-Reply-To:  <1133335040.960743.190080@g49g2000cwa.googlegroups.com>
Content-Type: text/plain; charset=ISO-8859-1

LXY,

if you have a regression model y = f(x1, x2). then the simple scatter plots
will be y*x1for x1 and y*x2 for x2 and partial plots will be y-f(x1)*x2 for
x2 and y-f(x2)*x1 for x1.

On 11/30/05, lixiaoyao <lixiaoyao5880@yahoo.com> wrote:
>
>   Hi all
>   I want to do a multiple regression,I read the link
> http://www.ats.ucla.edu/stat/sas/webbooks/reg/chapter2/sasreg2.htm,about
> how to test the nonlinearity,they show me a simple regression,and they
> said
> "
> This section has shown how you can use scatterplots to diagnose
> problems of non-linearity, both by looking at the scatterplots of the
> predictor and outcome variable, as well as by examining the residuals
> by predicted values.  These examples have focused on simple regression;
> however, similar techniques would be useful in multiple regression.
> However, when using multiple regression, it would be more useful to
> examine partial regression plots instead of the simple scatterplots
> between the predictor variables and the outcome variable.
>
> "
>
> I do not understand "what is the partial regression plots instead of
> the simple scatterplots between the predictor variables and the outcome
> variable"
>   Can anyone explain this more to me or give me a detail link?
>   Thank you for your time!
>



--
WenSui Liu
(http://statcompute.blogspot.com)
Senior Decision Support Analyst
Health Policy and Clinical Effectiveness
Cincinnati Children Hospital Medical Center