PROC MIXED LSMEANS

deleted_user — Wed, 19 Mar 2008 15:28:10 GMT

Let me preface this by saying I am a relative novice using PROC MIXED. I am using PROC MIXED to conduct a repeated measures ANOVA. I am using the following code:
proc mixed data=ds order=data;
class SUB IV;
model DV=IV;
repeated /subject=SUB type = cs r rcorr;
lsmeans IV;
run;

Where IV has 3 levels. Also, there is no missing data. When I run this I the standard errors produced by the LSMEANS statement are the same for all 3 levels. I'd like to understand why this is so. I am interested in reporting the LSMEANS and SE.

Thanks,
Jeff

Re: PROC MIXED LSMEANS

statsplank — Tue, 21 Oct 2008 04:32:14 GMT

Hi Jeff,

First, I would like to note that the way you wrote the PROC MIXED code corresponds to RCBD with subjects as blocks. If that is the case, then the standard errors for all tretament levels are computed using the same formula:

se[y_bar]=SQRT{(var1+var2)/n},

where var1 and var2 are, respectively, a between-subject variance (block variation) and a MS(Error), and n is a number of subjects.

topic PROC MIXED LSMEANS in SAS Health and Life Sciences

PROC MIXED LSMEANS

Re: PROC MIXED LSMEANS