Is there a particular way I need to define fixed and random variables for PROC GLM

[RFETT]

I am currently trying to analyze the differences between three measurement machines. Every time I attempt to run the code I get an error message saying the statement is out of  proper order or not valid. Machine is the fixed effect and Rep is the random effect. Here is the code I am currently using. Any assistance would be appreciated.

Proc GLM data= robertU4L;
Class Machine Rep;
Model TW = Machine Rep;
Fixed= Machine;
Random= Rep;
means machine/lsd;
Run;

[PaigeMiller]

There is no FIXED statement in PROC GLM, which assumes if the variable isn't in the RANDOM statement, then it is fixed.

For your reference, here is the PROC GLM documentation, which you should bookmark if you are going to be using PROC GLM a lot.

https://documentation.sas.com/doc/en/pgmmvacdc/9.4/statug/statug_glm_toc.htm

[PaigeMiller]

Depending on what you mean by REP you may have mis-specified the model.

If REP is truly a replicate number, then you would not want it in the model. On the other hand, if REP is something else, then it does belong in the model.