Degrees of Freedom in Proc Mixed with Repeated and Random Statements

AW_Carter — Wed, 26 Feb 2020 18:48:54 GMT

I am running a repeated measures in Proc Mixed using a repeated statement. I also have a random effect that I would like to include. When I run the model without the random effect, my den. df reflects the number of subjects (as it should). However, as soon as I add in a random statement, my df nearly triples. Why? Is there a way to include a random effect with a repeated measures that doesn't inflate my df?

Here is the code with both the random and repeated statements:

proc mixed Data=Flux_Devo method=ml;
class id family trt stage;
model CO2lograte= temp|trt|stage mass_start;

random family;
repeated / type=un subject=id r;
run;

Thank you!

topic Degrees of Freedom in Proc Mixed with Repeated and Random Statements in Statistical Procedures

Degrees of Freedom in Proc Mixed with Repeated and Random Statements