topic Re: Repeated Measure ANOVA on RAnks in Statistical Procedures

Repeated Measure ANOVA on RAnks

LucyB — Fri, 31 Mar 2017 03:35:55 GMT

I have some one group pre-post data and I am looking at differences between a group variable (binary) and the pre post change in the score outcome.

My code for the analysis is below:

proc glm data=data;
   class group;
   model pre post = group / nouni;
   repeated time 2 (1 2) / printe;
   lsmeans group / stderr;
run;

Ksharp — Thu, 30 Mar 2017 08:26:41 GMT

I would try PROC MIXED instead of PROC GLM.

Check GLM documenation to see how to transform Repeated Measure from GLM into MIXED.

change your data structure as

data have;

set x;

sub+1;

time=1;value=pre;output;

time=2;value=post;output;

run;

proc mixed data=have;

class group time;

model value=group;

repeated time/type=un solution subject=sub;

lsmean group/pdiff cl;

run;

LucyB — Thu, 30 Mar 2017 20:49:15 GMT

but does this still not make the same assumptions? my outcome is skewed.

Ksharp — Fri, 31 Mar 2017 12:52:52 GMT

If you are trying to do Repeated Measure,
I suggest you use PROC MIXED , not GLM .
GLM can't do the real mixed model .