Hello, I'm new to mixed models, having run essentially all of my data for my dissertation using GLM. Normally, we'd have chosen 12 mice families (strains = 6 are cancer-prone, 6 are not) and would measure gene expression as a function of time, family, strain, and the interaction of time and strain using:

CLASS TIME STRAIN FAMILY; MODEL p53 = TIME STRAIN FAMILY(STRAIN) TIME* STRAIN; LSMEANS TIME*STRAIN #contrast statements, etc. using the nested term to 'block' for families. There is a different gene that was unanticipated, and we have re-classified families into different STRAINS which requires a random model. My limited understanding of code would be:

CLASS TIME STRAIN FAMILY; MODEL p53 = TIME STRAIN TIME* STRAIN; RANDOM FAMILY(STRAIN);

Does this model accomplish testing the nested term over the error and properly treat the random variable? Suggestions or reading would be appreciated.

Thanks,

Maurice