Date: Thu, 11 Apr 2002 10:57:26 -0400
Reply-To: bramley.m@PG.COM
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: bramley.m@PG.COM
Subject: Re: regression line in GPLOT?
Content-type: text/plain; charset=us-ascii
All:
On the Symbol statement, there is a MODE=Exclude|Include option, and the default
is Exclude. So, add Mode=Include to fix this.
Sincerely,
Michael Bramley
Cha Gheill!
SAS - The Power To Know(tm)
SAS-L - The Power To Know More(tm)
Internet Mail Message
Received from host: mailgw.cc.uga.edu
[128.192.1.101]
From: Paige Miller <paige.miller@KODAK.COM> on 04/11/2002 02:13 PM GMT
Paige Miller To: SAS-L@LISTSERV.UGA.EDU
<paige.miller@KODAK.COM> Cc: (bcc: Michael Bramley-M/PGI)
Subject: Re: regression line in GPLOT?
04/11/2002 10:13 AM
Please respond to Paige
Miller
<paige.miller@KODAK.COM>
jerry chiu wrote:
>
> I have a proc gplot with INTERPOL=RLCLM95
> and proc reg, (see the code below)
> but they gave me very different regression lines,
>
> wonder what did I do wrong,
> Thanks,
> Jerry
>
> proc gplot data=plotdata ;
> plot height * weight ;
> ;
>
> BY group bydos2w ;
> run ;
>
> symbol1 interpol=rLclm95
> value=star
> height =1
> cv=red
> ci=blue
> co=green
> width=2;
>
> plot height * weight
> / vaxis=-50 to 150 by 10
> haxis=-4 to 4
> vminor=1
> regeqn;
> run;
>
> proc REG data=plotdata ;
> model height = weight ;
> by group bydos2 ;
I am guessing that the problem is your vaxis and haxis options. While
they might make the plot more attractive, they also cause PROC GPLOT to
ignore certain data points, and not use them in the regression if the
data point is beyond the limits of your haxis and vaxis.
--
Paige Miller
pmiller5@rochester.rr.com
"It's nothing until I call it!" -- Bill Klem, NL Umpire
"When you get the choice to sit it out or dance,
I hope you dance" -- Lee Ann Womack
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.