LISTSERV at the University of Georgia
Menubar Imagemap
Home Browse Manage Request Manuals Register
Previous messageNext messagePrevious in topicNext in topicPrevious by same authorNext by same authorPrevious page (November 2008, week 3)Back to main SAS-L pageJoin or leave SAS-L (or change settings)ReplyPost a new messageSearchProportional fontNon-proportional font
Date:         Mon, 17 Nov 2008 16:41:09 -0600
Reply-To:     Robin R High <rhigh@UNMC.EDU>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Robin R High <rhigh@UNMC.EDU>
Subject:      Re: Proc Mixed help
Comments: To: Brad Heins <hein0106@UMN.EDU>
In-Reply-To:  <200811172007.mAHGhL20025477@malibu.cc.uga.edu>
Content-Type: text/plain; charset="US-ASCII"

Brad,

Work through this example dataset with each MODEL statement run in turn may help demonstrate how you can dummy code the continuous variable of interest.

DATA tst; cls=1; do x = 1 to 15; x1=x; x2=0; x3=0; y = 5 + 2.5*x + 1.2*rannor(929); OUTPUT; END; cls=2; do x = 1 to 15; x1=0; x2=x; x3=0; y = 7 + .10*x + 1.2*rannor(0); OUTPUT; END; cls=3; do x = 1 to 15; x1=0; x2=0; x3=x; y = 8 + .15*x + 1.2*rannor(0); OUTPUT; END;

ods select solutionF;

PROC MIXED; CLASS cls; MODEL Y = cls x(cls) / solution ; * MODEL y = cls x1 x2 x3/ solution ; * MODEL y = cls x1 / solution ; run;

Robin High UNMC

Brad Heins <hein0106@UMN.EDU> Sent by: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU> 11/17/2008 02:10 PM Please respond to Brad Heins <hein0106@UMN.EDU>

To SAS-L@LISTSERV.UGA.EDU cc

Subject Proc Mixed help

I have a question about proc mixed at if there is a trick I can do with my model statement or not.

I have a variable aocm that is continuous and is nested within a class variable (LNR) that has 3 levels. From the output below you can see that at LNR 2 and 3 the regression coefficient is not significant, but it for LNR 1.

Is there a way that I can tell Proc mixed to just adjust for only LNR 1 and not for LNR 2 or 3 since they are not significant. I want to leave LNR 1 is the model and adjust for those records, but not for LNR 2 or 3 because the regression coefficients make no biological sense to the data that I have.

Any help would be appreciated. Thanks. Brad

proc mixed statement: class group lnr hy; model y= group lnr hy aocm(lnr) ;

Effect Estimate Error DF t Value Pr > |t| aocm(LNR) 1 2.1697 0.2378 9660 4.92 <.0001 aocm(LNR) 2 -1.0259 0.2432 9660 8.33 <.6538 aocm(LNR) 3 -1.4974 0.2572 9660 5.82 <.8520


Back to: Top of message | Previous page | Main SAS-L page